In clinical practice, much time is spent on measuring the clinical parameters of glaucoma including the intra-ocular pressure (IOP), visual acuity (VA), visual field, and retinal nerve fibre layer (RNFL) thickness. What is often neglected is the quality of life (QOL) of patients and how well they live with their disease on a day-to-day basis. Glaucoma affects 80 million people worldwide.1 It is a chronic and irreversible disease with a heavy burden on visual function and vision, besides being one of the most important constituents affecting QOL.2 3 4

Recourse to QOL questionnaires in glaucoma can be broadly divided into general health–related, vision-specific, or glaucoma-specific.5 Quality-of-life assessment in glaucoma patients is as important as the clinical parameters used to measure glaucoma progression, because it reflects the impact of the ocular disease on the patient as a whole and may also be an indicator of whether the disease is advancing.4 6 7 8 9

Using generic QOL assessments, glaucoma was found to have deleterious impact as other systemic chronic diseases like osteoporosis, diabetes, or dementia.10 However, such generic tests do not address the end points of glaucoma, such as visual impairment and visual field constriction, for which reason their robustness and specificity are limited.10 There are approximately 18 different patient-reported QOL assessments specific to glaucoma. Among these, the Glaucoma Quality of Life–15 Questionnaire (GQL-15) and the Vision and Quality of Life Index have been found most satisfactory in terms of content, validity, and reliability.11 Thus, the aim of this study was to investigate the correlations between clinical parameters and glaucoma-specific QOL in Chinese patients with bilateral primary open-angle glaucoma (POAG).

For this cross-sectional study, consecutive patients with bilateral POAG were recruited from an academic hospital in Hong Kong. The diagnosis of POAG was based on an open angle on gonioscopy, a presenting IOP of >21 mm Hg, and either a glaucomatous visual field loss on at least two Humphrey visual field tracings using the 24-2 SITA fast protocol (Humphrey Instruments, Inc, Zeiss Humphrey, San Leandro [CA], US) or RNFL thinning on Spectralis Optical Coherence Tomography (Heidelberg Engineering, Carlsbad [CA], US). Patients were excluded if they had unilateral disease, concomitant ocular diseases that significantly affected their vision (amblyopia, mature cataract affecting the accuracy of glaucoma investigations). Patients were also excluded if they had other corneal or retinal pathologies, or if they were unable to yield reliable visual field results. Their IOPs were determined using Goldmann applanation tonometry.

The GQL-15 questionnaire is glaucoma-specific, and assesses patient-perceived visual disability in 15 daily tasks responded to in writing. The tasks addressed four aspects of visual disability: (1) central and near vision; (2) peripheral vision; (3) dark adaptation and glare; and (4) outdoor mobility. A 5-point rating scale for the level of difficulty of each task can yield a total score of 0 to 75. Higher scores signify a lower QOL. The GQL-15 was translated into traditional Chinese text and distributed to participating patients. For illiterate patients, the items were read out to them in Cantonese dialect. The questionnaire was translated from English to Chinese by an investigator who was fluent in both English and Chinese. The translated questionnaire was checked for discrepancies by a second investigator and a consensus was reached to develop a draft Chinese questionnaire. A third investigator then back-translated the draft Chinese questionnaire into English; the back-translated draft and the original version were then compared. Discrepancies were amended and gave rise to the final Chinese version. The questionnaire was then tested on five POAG patients of varying gender and age. Patients were asked to complete the questionnaire, and offer their own interpretation of its contents and whether any alternative wording should be used.

The D’Agostino-Pearson omnibus test was used to test for normality. Nearly half of the parameters passed the normality testing. The means of several clinical parameters were calculated for the two eyes and correlated with the GQL-15 using Pearson’s correlation coefficient and linear regression analysis. The selected parameters were the visual field index (VFI) and pattern standard deviation (PSD) from the Humphrey Field Analyzer, the Snellen best-corrected VA, the presenting IOP, current IOP, average RNFL thickness via optical coherence tomography, as well as the number of topical anti-glaucoma medications being used. t Tests were used to test for differences between the mean GQL-15 scores between males and females. Data were expressed as mean ± standard deviation (SD). Any P value of <0.05 was accepted as statistically significant.

Our institutional review board granted ethics approval for the study and informed consent was obtained from each patient prior to the start of the study.

Fifty-one patients with bilateral POAG were recruited, all of whom were Chinese. Their mean (± SD) age was 65.8 ± 12.1 years and the male-to-female ratio was 1.1:1.

The means of their clinical parameters for both eyes are shown in the Table. Their mean GQL-15 score was 26.0 ± 11.6 (out of 75). The three most problematic activities reported for all patients belonged to: item 4 “adjusting to bright lights” (mean score, 2.3 ± 1.3); item 6 “going from a light to a dark room or vice versa” (mean score, 2.3 ± 1.3); and item 2 “seeing at night” (mean score, 2.2 ± 1.2).

There was a moderately significant correlation between a lower VFI and a poorer GQL-15 score (r=0.3, P=0.01; Fig 1). Likewise, a poorer VA correlated significantly with a poorer GQL-15 score (r=0.3, P=0.03; Fig 2). These two correlations seemed to follow a linear pattern such that linear regression analysis showed a weak linear relationship between a poorer GQL-15 score and a lower VFI (r²=0.1, P=0.01) and a poorer VA (r²=0.1, P=0.03).

A thinner RNFL appeared to be associated with a poorer GQL-15 score but the correlation did not attain statistical significance (r=0.3, P=0.07). In terms of pressure control, a higher presenting IOP showed a trend towards correlation with a poorer GQL-15 score (r=0.2) as did a lower current IOP (r= 0.2) and a greater number of anti-glaucoma eye drops used (r=0.1). However, none of these correlations reached statistical significance (all P>0.7). On comparing GQL-15 scores between male and female glaucoma patients, no significant difference was found (P=0.3, t test).

Various studies have associated QOL with visual field impairment.8 12 Odberg et al13 simply categorised visual field defects into “normal”, “having a restricted scotoma”, or “having a field defect large enough to be of visual significance”, and found a weak-to-moderate correlation between such visual field defects and subjective visual disabilities. The Collaborative Initial Glaucoma Treatment Study later found that at the time of diagnosis, patients’ visual fields correlated only modestly with a health-related QOL questionnaire and that of VFIs; mean deviation (MD) showed better correlation with QOL than PSD, corrected pattern SD, or short-term fluctuation.14 Nelson et al4 found that the GQL-15 scores, and especially the subsets pertaining to glare, correlated significantly with MD, even for patients with mild disease. Furthermore, those with moderate and severe visual field loss had similar GQL-15 scores, suggesting a threshold for disability may be reached up to a certain level of glaucoma severity4 or represent adaptation to loss of visual function. Similarly, Goldberg et al15 have found that the GQL-15 scores correlated with VA, MD, the number of binocular points of <10 dB, and that QOL tended to decrease with disease severity. Whilst MD is commonly correlated with QOL in glaucoma patients, it has the drawback of not being specific enough to represent the limitations caused by glaucoma alone, since it may also be affected by global defects like cataract. On the other hand, using PSD eliminates the factor of global defects, though it is not sensitive in advanced glaucoma, where the entire field is globally depressed.

Thus in this study, we utilised the VFI, which is a percentage summarising the overall visual field status compared to age-adjusted visual fields. The VFI emphasises the importance of the central field. It is less affected by media opacities (cataracts), and is more accurate than MD for monitoring glaucoma progression.16 17 Few studies have used VFI to correlate with QOL in glaucoma. Sawada et al18 reported that VFI correlated with QOL via the 25-item National Eye Institute Visual Function Questionnaire (NEI VFQ-25) and that the correlation was better than with MD. Our study found a statistically significant correlation between the reduction in mean binocular VFI and a poorer GQL-15 score and that VFI was a better indicator of glaucoma-specific QOL than RNFL thickness, IOP, or PSD on visual field. We chose to use PSD rather than MD in our analysis because the latter could be affected by any global obstruction to vision like cataract, whereas PSD is more specific for inter-field variability. However, the two clinical parameters that achieved a significant correlation with the GQL-15 score were binocular VFI and VA, and both parameters were also associated with the GQL-15 score in a linear manner.

Intra-ocular pressure control did not correlate significantly with QOL although a higher IOP on presentation seemed to produce a lower QOL score, and interestingly a lower current IOP seemed to correlate with a poorer QOL. This unique finding may indicate that those with a lower current IOP have had glaucoma for longer or have more advanced disease warranting more aggressive pressure reduction. Furthermore, those using more anti-glaucoma eye drops seemed to have a lower QOL score, but these correlations were weak and did not reach statistical significance.

Patient perceptions of disease and methods of coping are heavily influenced by culture and ethnicity. Thus, Singapore Chinese glaucoma patients were more accepting of their daily disabilities than corresponding American Caucasians.19 Literature pertaining to Chinese glaucoma patients is sparse. Wu et al20 found that Chinese glaucoma patients were particularly concerned about the uncertainties of treatment, the prognosis, and passing on of the disease to family members. Lin and Yang21 reported a correlation with MD and the Medical Outcomes Study Short-Form 36 Health Survey and the NEI VFQ-25. Whilst clinical data provide evidence of structural and functional damage of the optic nerve, they do not address the impact of disease on patients. The correlation of objective clinical measurements to QOL is particularly useful, because it gives ophthalmologists in a busy clinical setting an overall impression of glaucoma-specific QOL. This can enable them to recommend environmental and lifestyle modifications to minimise obstacles and maximise the period of independence.5 Our study found that in Chinese glaucoma patients, the most problematic aspects of coping were “adjusting to bright lights”, “going from a light to a dark room or vice versa”, and “seeing at night”. Interestingly, all these activities belong to the realm of dark adaptation. Hence, environmental modifications can potentially help to reduce glare.4 Furthermore, an estimation of QOL from clinical parameters can allow ophthalmologists to more readily identify patients with a poorer QOL needing more psychosocial support. Interestingly, it has been reported that POAG itself is associated with anxiety, depression, and hypochrondriasis22 and a low GQL-15 score has also been identified as a predictor for depression.23

One limitation of our study was that it was cross-sectional and looked at POAG patients with varying degrees of severity. A longitudinal study would have provided additional information about the changes in QOL throughout different stages of the disease. A second limitation was that the population received heterogeneous treatments (lasers and surgeries). However, as the aim of this study did not involve evaluating the side-effects of glaucoma treatments and since the GQL-15 too did not target treatment side-effects, we did not consider it necessary to exclude those who had undergone such treatments previously. Rather, we opted to include a more heterogeneous POAG population to make the results more generalisable and representative. A third limitation was that no single test is perfect; the GQL-15 mainly focuses on visual activities, which is only one aspect of QOL. Conceivably, such a questionnaire only reflects patient confidence to perform certain tasks rather than the actual difficulties experienced. Nevertheless, it has been shown that patients’ loss of confidence often precedes their perceptions of difficulty.24

To the best of our knowledge, this is one of the few studies reporting a significant correlation and a linear relationship between VFI and the glaucoma-specific GQL-15 score in the Chinese POAG patients. This study also identified dark adaptation as the most challenging visual issue pertinent to Chinese POAG patients.

No conflicts of interest were declared by the authors.

1. Mansberger SL, Demirel S. Early detection of glaucomatous visual field loss: why, what, where, and how. Ophthalmol Clin North Am 2005;18:365-73, v-vi. CrossRef

2. Beauchamp CL, Beauchamp GR, Stager DR Sr, Brown MM, Brown GC, Felius J. The cost utility of strabismus surgery in adults. J AAPOS 2006;10:394-9. CrossRef

3. Brown GC, Brown MM, Sharma S, et al. The burden of age-related macular degeneration: a value-based medicine analysis. Trans Am Ophthalmol Soc 2005;103:173-86.

4. Nelson P, Aspinall P, Papasouliotis O, Worton B, O’Brien C. Quality of life in glaucoma and its relationship with visual function. J Glaucoma 2003;12:139-50. CrossRef

5. Spaeth G, Walt J, Keener J. Evaluation of quality of life for patients with glaucoma. Am J Ophthalmol 2006;141(1 Suppl):S3-14. CrossRef

6. Jampel HD, Schwartz A, Pollack I, Abrams D, Weiss H, Miller R. Glaucoma patients' assessment of their visual function and quality of life. J Glaucoma 2002;11:154-63. CrossRef

7. Janz NK, Wren PA, Lichter PR, Musch DC, Gillespie BW, Guire KE. Quality of life in newly diagnosed glaucoma patients: the Collaborative Initial Glaucoma Treatment Study. Ophthalmology 2002;108:887-97; discussion 898. CrossRef

8. Parrish RK 2nd, Gedde SJ, Scott IU, et al. Visual function and quality of life among patients with glaucoma. Arch Ophthalmol 1997;115:1447-55. CrossRef

9. Gutierrez P, Wilson MR, Johnson C, et al. Influence of glaucomatous visual field loss on health-related quality of life. Arch Ophthalmol 1997;115:777-84. CrossRef

10. Mills T, Law SK, Walt J, Buchholz P, Hansen J. Quality of life in glaucoma and three other chronic diseases: a systematic literature review. Drugs Aging 2009;26:933-50. CrossRef

11. Vandenbroeck S, De Geest S, Zeyen T, Stalmans I, Dobbels F. Patient-reported outcomes (PRO's) in glaucoma: a systematic review. Eye (Lond) 2011;25:555-77. CrossRef

12. Lee BL, Gutierrez P, Gordon M, et al. The Glaucoma Symptom Scale. A brief index of glaucoma-specific symptoms. Arch Ophthalmol 1998;16:861-6. CrossRef

13. Odberg T, Jakobsen JE, Hultgren SJ, Halseide R. The impact of glaucoma on the quality of life of patients in Norway. II. Patient response correlated to objective data. Acta Ophthalmol Scand 2001;79:121-4. CrossRef

14. Mills RP, Janz NK, Wren PA, Guire KE. Correlation of visual field with quality-of-life measures at diagnosis in the Collaborative Initial Glaucoma Treatment Study (CIGTS). J Glaucoma 2001;10:192-8. CrossRef

15. Goldberg I, Clement CI, Chiang TH, et al. Assessing quality of life in patients with glaucoma using the Glaucoma Quality of Life–15 (GQL-15) questionnaire. J Glaucoma 2009;18:6-12. CrossRef

16. Bengtsson B, Heijl A. A visual field index for calculation of glaucoma rate of progression. Am J Ophthalmol 2008;145:343-53. CrossRef

17. Casas-Llera P, Rebolleda G, Muñoz-Negrete FJ, Arnalich-Montiel F, Pérez-López M, Fernández-Buenaga R. Visual field index rate and event-based glaucoma progression analysis: comparison in a glaucoma population. Br J Ophthalmol 2009;93:1576-9. CrossRef

18. Sawada H, Fukuchi T, Abe H. Evaluation of the relationship between quality of vision and the visual function index in Japanese glaucoma patients. Graefes Arch Clin Exp Ophthalmol 2011;249:1721-7. CrossRef

19. Saw SM, Gazzard G, Au Eong KG, Oen F, Seah S. Utility values in Singapore Chinese adults with primary open-angle and primary angle-closure glaucoma. J Glaucoma 2005;14:455-62. CrossRef

20. Wu PX, Guo WY, Xia HO, Lu HJ, Xi SX. Patients' experience of living with glaucoma: a phenomenological study. J Adv Nurs 2011;67:800-10. CrossRef

21. Lin JC, Yang MC. Correlation of visual function with health-related quality of life in glaucoma patients. J Eval Clin Pract 2010;16:134-40. CrossRef

22. Erb C, Thiel HJ, Flammer J. The psychology of the glaucoma patient. Curr Opin Ophthalmol 1998;9:65-70. CrossRef

23. Skalicky S, Goldberg I. Depression and quality of life in patients with glaucoma: a cross-sectional analysis using the Geriatric Depression Scale–15, assessment of function related to vision, and the Glaucoma Quality of Life–15. J Glaucoma 2008;17:546-51. CrossRef

24. Nelson P, Aspinall P, O’Brien C. Patients’ perception of visual impairment in glaucoma: a pilot study. Br J Ophthalmol 1999;83:546-52. CrossRef

Type 2 diabetes mellitus (T2DM) is one of the most common chronic conditions encountered in primary care, and affects up to 10% of Hong Kong (HK) population.1 Its complications include kidney disease, blindness, lower limb amputation, and coronary heart disease; all of which lead to increased morbidity and mortality.2

Ethnic minorities constitute an important component of the HK population. According to census in 2011, about 95% of the local inhabitants are ethnic Chinese; the remainder (ethnic minorities) are mainly from Asia (India, Philippines, Nepal, Pakistan, and Indonesia).3 Previous studies have shown that diabetes affects certain ethnic minority groups (EMGs) differently.4 South Asians are at higher risk for T2DM by up to 4 to 6 fold compared with other ethnic groups, probably due to a combination of genetic and environmental factors.5 6 In addition, South Asians have a much higher prevalence of T2DM with cardiovascular disease that occurs at an earlier age and is associated with higher morbidity and mortality.7 Differences in health care systems, limited access to health services, and social deprivation can further compound the risk of developing diabetes and its complications.

Improving the quality of chronic disease management is an essential component of health policy in the community. Locally, a significant proportion of T2DM patients including those from EMGs are managed in primary care and followed up at government general out-patient clinics (GOPCs) of the Hong Kong Hospital Authority (HKHA). The clinic where the authors work is one of the largest GOPCs of the HKHA, and more than 50% of its attendees have chronic diseases including diabetes. In addition, it is located in central Kowloon, where most of the South Asian minorities including Indians, Nepalese, and Pakistanis reside.

Till now, local data on the diabetic control among EMG diabetes patients are lacking. To address this knowledge gap, we aimed to identify and compare the demographics of diabetes and its control in ethnic minority and Chinese patients managed in primary care and to explore possible strategies to improve care. We believe this study will provide important background information to address important issues pertinent to chronic disease management within various HK ethnic groups.

This was a retrospective case series study carried out in the Yau Ma Tei Jockey Club GOPC of the HKHA. According to a pilot study carried out in early 2012, the five major ethnic minorities undergoing regular follow-up in this clinic were from India, Nepal, the Philippines, Pakistan, and Indonesia. Regular follow-up was defined as returning to our clinic for chronic disease management on a regular basis, ie, every 1 to 4 months. Very few Caucasians or other Asian ethnic groups such as the Japanese and Koreans had regular follow-up at this clinic and were therefore excluded from the analysis.

Patients with T2DM coded by International Classification of Primary Care (ICPC) T90, who had been regularly followed up at Yau Ma Tei Jockey Club Clinic between 1 March 2012 and 28 February 2013 and had an annual blood and urine checkup at least once during this period, were recruited. The diagnosis of diabetes was based on the “Definition and description of diabetes mellitus” from American Diabetes Association in 2010.8 Wrongly diagnosed diabetes patients, type 1 diabetes patients, diabetes patients who were regularly followed up in the specialist out-patient departments (SOPDs), diabetes patients who had no annual checkup within this period, and those who were neither Chinese nor belonged to the above five EMGs were excluded.

The recruited patients’ age, gender, ethnicity, smoking status, body mass index (BMI), latest blood pressure, fasting blood sugar (FBS), haemoglobin A1c (HbA1c) and creatinine levels, urine albumin/creatinine ratio, and lipid profile were retrieved from the Clinical Management System (CMS) of the HKHA. The most recent blood and urine test was used for analysis if more than one test had been performed during the study period. The BMI was calculated as body weight/body height²(kg/m²).The patient was considered a smoker if he/she currently smoked or was in the first 6 months of stopping.

We used the abbreviated Modification of Diet in Renal Disease9 to give an estimated glomerular filtration rate (eGFR) expressed in mL/min/1.73 m², and chronic kidney disease was defined as having an eGFR of <60 mL/min/1.73 m²:
eGFR=186 × [SCR/88.4]^–1.154 × [age]^–0.203 × [0.742 if female]
where SCR was the serum creatinine level expressed as µmol/L

The medical history of stroke, ischaemic heart disease (IHD), and concomitant hypertension (HT) were retrieved based on ICPC codes in the CMS. Stroke cases were retrieved using ICPC codes K89 (transient ischaemic attack), K90 (cerebrovascular accident), and K91 (cerebrovascular disease). Cases of HT were retrieved using ICPC codes K86 (uncomplicated HT) and K87 (complicated HT). Patients with IHD were retrieved using the codes K74 (IHD with angina), K75 (acute myocardial infarction), and K76 (IHD without angina). Repeat systolic blood pressures (SBPs) of ≥130 mm Hg or diastolic blood pressures (DBPs) of ≥80 mm Hg confirmed a diagnosis of HT in diabetes patients.10

All data were entered and analysed using computer software (Statistical Package for the Social Sciences; Windows version 16.0; SPSS Inc, Chicago [IL], US). Student’s t test and analysis of variance were used to analyse continuous variables and Chi squared tests for categorical data. Tukey and Games-Howell tests were used for pairwise comparisons within the five minority groups, if applicable. All statistical tests were two-sided, and a P value of <0.05 was considered significant.

A list of 5536 T2DM patients followed up in this clinic from 1 March 2012 to 28 February 2013 was generated from the CMS. Among them, 1190 (21.5%) were excluded due to the already described exclusion criteria (11 wrongly diagnosed as diabetic, 1 had type 1 diabetes, 395 were regularly followed up in the SOPDs, 2 were Caucasians, and 781 diabetes patients had no blood and urine check-up during the recruitment period). Thus, findings from the remaining 4346 (78.5%) patients fulfilling our inclusion criteria were analysed. Among these patients, 3966 (91.3%) were Chinese and 380 (8.7%) were from the EMGs. Table 1 summarises the demographic characteristics of these patients in both the Chinese and EMGs. In summary, they were comparable in terms of gender ratio and smoking status (both P>0.05). However, patients from the EMGs were significantly younger (mean ± standard deviation [SD], 55.4 ± 11.7 years vs 66.1 ± 11.5 years; P<0.001) and their BMIs were much higher (mean ± SD, 28.5 ± 4.6 kg/m² vs 25.8 ± 4.3 kg/m²; P<0.001) than those of the Chinese diabetes patients.

To reduce confounding due to age, 380 age- and sex-matched diabetes patients were randomly selected from the Chinese diabetes cohort. Table 2 summarises the glycaemic, blood pressure and lipid profile control, as well as kidney function in these diabetic Chinese and EMGs. The latter patients were found to have a greater proportion with HT than the Chinese diabetic controls (P=0.03), whereas their co-morbidity rates for stroke, IHD, and chronic kidney disease were similar. Glycaemic control was poorer in EMG diabetes patients than their age- and sex-matched Chinese counterparts (HbA1c, 7.8 ± 1.7% vs 7.5 ± 1.4%; P=0.006). Consistently, their FBS levels were also much higher than those of the controls (P=0.02). With regard to blood pressure control, SBP was similar in the two groups, but the mean DBP was higher in the EMG cohort (78 ± 11 vs 73 ± 11 mm Hg; P<0.001). When lipid control was compared, total cholesterol, low-density lipoprotein (LDL), and triglyceride levels were found to be similar in the two groups. High-density lipoprotein levels (HDLs), however, were much lower in the EMG diabetes patients (1.19 ± 0.33 mmol/L vs 1.28 ± 0.36 mmol/L; P=0.001).

Regarding the demographic characteristics of EMG diabetes patients (Table 3), most were Nepalese (n=169), followed by Indian (n=103), Filipino (n=51), Pakistani (n=47), and Indonesian (n=10). The male-to-female (M/F) ratio was much higher in the Pakistani, Indian, and Nepalese groups (P<0.001). However, the mean age of the Nepalese and Pakistani patients was much younger than that of the Indian and Indonesian groups (P=0.004). More Nepalese and Pakistani diabetes patients were chronic smokers than those from the other ethnic minorities (P<0.001).

Table 4 shows glycaemic, blood pressure, and lipid profile control in diabetes patients within the individual EMGs. Owing to their dissimilar age and gender composition, comparisons between different minority groups were inevitably confounded. Nevertheless, the data indicated that glycaemic control was particularly poor in Pakistani patients (mean ± SD HbA1c levels being 8.4 ± 1.6%), and less so in the Nepalese and Indian groups (7.8 ± 1.9% and 7.8 ± 1.7%, respectively). In contrast, the metabolic control of Indonesian diabetes patients was generally satisfactory (mean HbA1c level being 6.8 ± 0.6%). The mean SBP was similar among all EMGs, but the mean DBP control was suboptimal in the Nepalese group (84 ± 11 mm Hg) and within target in the other minority groups. When lipid control was studied, the total cholesterol, LDL, and triglyceride levels were similar, but Pakistani patients had a much lower mean HDL level (1.04 ± 0.27 mmol/L).

This study was the first clinical analysis of T2DM patients in local EMGs. It compared demographic characteristics of both Chinese and EMG diabetes patients managed in primary care. Notably, it revealed discrepancies between the groups in terms of glycaemic, blood pressure, and lipid profile control.

Notably, in HK, the basic demographic features of Chinese diabetes patients and those from EMGs were quite different. The latter were younger and more obese; such findings were in line with those in the HK census in 2011 which showed that 61.3% of EMGs were aged 25 to 44 years and that the median age for all EMG patients was much lower than that of the entire HK population.3 In addition, the main reason for staying in HK for nearly all EMG subjects was to work, and when asked about their occupation most of the recruited EMG diabetes patients (n=334, or 87.9% of them) stated that they undertook manual labour. Thus, most were in their 40s and 50s and therefore their mean age was understandably younger than that of their Chinese counterparts (identified within a gradually ageing population). Furthermore, diabetes patients from South Asian ethnicities were more obese and had a much higher BMI than their Chinese controls. It is well known that the prevalence of obesity varies substantially between ethnic groups and is estimated to differ according to the precise measurements used (eg BMI, waist-to-hip ratio, and waist circumference). Although no data in the literature have directly compared the BMI of Chinese diabetes patients with that of those from South Asia, studies from UK have revealed that the mean waist-hip girth ratios and trunk skin folds were larger in South Asians than in European and Chinese groups.11

Since age is a very important confounder that prevented direct comparison between the two groups, age- and sex-matched diabetes patients from the Chinese and ethnic minorities were studied further. Even so, glycaemic control was poorer in EMG patients than the matched Chinese controls (mean ± SD, HbA1c 7.8 ± 1.7% vs 7.5 ± 1.4%; P=0.006). Whereas SBP control was similar, the mean DBP was higher in the EMGs (P<0.001). In addition, the mean HDL levels were much lower in EMGs than in the matched Chinese controls (P=0.001). Possible reasons for such a difference between could be multi-factorial. First, several studies have shown that genetic factors may play a determinant role.12 13 Diabetes patients from the South Asia appear more likely to have insulin resistance and a higher prevalence of obesity and metabolic syndrome, all of which are chronic conditions that challenge glucose metabolism.5 Second, patients from EMGs are often at a socio-economic disadvantage and difficult to reach via mainstream channels, and so they face inequalities in accessing medical care.3 For example, EMG diabetes patients might not have their diabetes diagnosed if they were socially disadvantaged and might be less inclined to seek medical care. Moreover, underdiagnosed individuals may be more likely to have poor diabetic control and experience early mortality. Third, the first language of South Asian groups is usually neither English nor Chinese, and therefore they may not understand the medical advice properly. Lastly, their cultures, religious beliefs, and lifestyles may influence their behaviour (including levels of physical activity and food choices), all of which affect health status and management. Coordinated efforts are therefore needed to overcome these limitations and embark on integrated diabetes monitoring and surveillance programmes in such EMGs.

We also need to be aware that a large proportion of diabetes patients followed up at public GOPCs are from lower-income groups and the geriatric populations. Younger Chinese T2DM patients might be more inclined to seek help from Specialist Clinics and private doctors. Thus, these findings might not be directly applicable to private or other specialist settings. Nevertheless, the present findings suggest important groundwork for further local and international studies.

The demographic characteristics of diabetes patients within EMGs indicated that their gender ratios also varied dramatically. Among Filipino patients, the M/F ratio was 0.31 and all Indonesian patients were female. By contrast, most Pakistani, Indian, and Nepalese diabetes patients were male (M/F ratios being 2.62, 1.51, and 1.35, respectively). These findings were consistent with a thematic report on ethnic minorities in the 2011 HK population consensus, which showed considerable variations in the gender composition of different ethnic groups in the community3; the M/F ratios of Indonesians and Filipinos were extremely low but the ratios were converse among Pakistanis and Nepalese. This was because large proportions of Filipinos and Indonesians in HK were foreign domestic helpers, of whom 99% were female.3 On the contrary, most Nepalese and Pakistanis worked in elementary occupations such as at construction sites or as security guards, and most were males. This difference in gender composition also contributed to a greater proportion of Nepalese and Pakistanis being chronic smokers as compared with the other Asian minorities. As the different age and sex distributions among EMGs was an important confounder of clinical outcomes, no direct comparison on diabetes control between different subgroups was feasible. Nevertheless, we found that Pakistani diabetes patients had particularly higher HbA1c levels and lower HDL concentrations. Indeed, studies have shown that the epidemiology and determinants of diabetes in Pakistan reveal a peculiar combination of risk factors.13 Strong genetic and environment factors interplay along with in-utero programming, in the context of low birth weights and gestational diabetes contributing to a high prevalence and poor control of T2DM in Pakistanis.14 On the other hand, Nepalese diabetes patients had suboptimal DBP control. This finding is in line with World Health Organization reports that Nepal has a high burden of HT and that the blood pressure control rates have been poor due to the inadequate awareness and lack of proper treatment.15 16 Local doctors should therefore pay particular attention to the needs of different ethnic groups and offer a flexible care package that reflects their physical, psychological, social, and cultural needs and at the same time upholds their autonomy, dignity, privacy, and personal choice.

Diabetes is a significant problem among both the Chinese and EMGs in HK. It is important that government officials, clinicians, and allied health workers understand the evidence and implement strategies to address shortcomings actively. Our local practice has emphasised empowering people with diabetes to support their own care management by proper diet control and active lifestyle strategies. In addition, concerted efforts are needed to raise awareness of diabetes and disseminate prevention messages to high-risk groups in collaboration with their community opinion leaders. Nowadays, information, interpretation, and advocacy services have been provided in HKHA clinics, which is definitely a positive step towards improving understanding of the disease among ethnic minority patients. Meanwhile, our services should assimilate aspects of ethnicity and culture, and implement culturally specific interventions to improve diabetes control in HK EMGs.

Family physicians are at the forefront of T2DM management, and aim to achieve optimal metabolic control to prevent macro- and micro-vascular complications. This study provides important background information on the demographic characteristics of diabetes patients from certain EMGs as compared to Chinese diabetes patients. Since certain South Asian groups tend to have poorer glycaemic control, culturally tailored health care interventions are required to improve their general health and chronic disease management.

One limitation was that only diabetes patients who were regularly followed up in a single clinic and had annual blood and urine checkups were studied. Second, the ethnic composition in other clinics and elsewhere in HK might differ considerably. Third, patients who were followed up at this clinic but never attended for annual assessment (n=781, 14.1%), whatever the reason, were excluded and must have given rise to a selection bias. However, we have compared the major epidemiological characteristics including age and gender of such patients and found that there were no obvious differences between them and the studied patients (P=0.45 and P=0.60, respectively). Fourth, all variables were measured at least once during the 1-year study period, and if more than one blood test was performed, the most recent result was used for analysis. Therefore, variability of measurements might have confounded the findings. Fifth, the relatively small sample size of certain EMG subgroups and their age and gender distribution discrepancies prevented direct comparison of their metabolic control. Nevertheless, the present results may lay the groundwork for similar studies in the future both locally and internationally. Lastly, concomitant chronic diseases (HT, IHD, and stroke) were retrieved via the ICPC code in the CMS, and so inadequate ICPC coding may have underestimated co-morbidity rates in both Chinese and EMG diabetes patients.

Ethnic minority groups are an integral part of the HK population. Compared with Chinese diabetes patients, EMG diabetes patients were much younger and more obese. Deficiencies existed in their understanding of diabetes management, particularly glycaemic control. Culturally tailored health care interventions are therefore necessary to promote patient education and clinical effectiveness for these patient groups and improve their long-term health.

We extend our gratitude to Dr King Chan for his continuous inspiration and support during this study. We also thank Ms Elise Chan, EA III of Department of Family Medicine and GOPC, for her patience during data entry and Mr Carl Chak, statistical officer of Queen Elisabeth Hospital, for his expertise and support in data analysis.

1. Chan JC, Malik V, Jia W, et al. Diabetes in Asia: epidemiology, risk factors, and pathophysiology. JAMA 2009;301:2129-40. CrossRef

2. Leung GM, Lam KS. Diabetic complications and their implications on health care in Asia. Hong Kong Med J 2000;6:61-8.

3. Hong Kong 2011 population census thematic report: ethnic minorities. Available from: http://www.statistics.gov.hk/pub/B11200622012XXXXB0100.pdf. Accessed Dec 2012.

4. Abate N, Chandalia M. The impact of ethnicity on type 2 diabetes. J Diabetes Complications 2003;17:39-58. CrossRef

5. McKeigue PM, Shah B, Marmot MG. Relation of central obesity and insulin resistance with high diabetes prevalence and cardiovascular risk in South Asians. Lancet 1991;337:382-6. CrossRef

6. Khan NA, Wang H, Anand S, et al. Ethnicity and sex affect diabetes incidence and outcomes. Diabetes Care 2011;34:96-101. CrossRef

7. Gholap N, Davies M, Patel K, Sattar N, Khunti K. Type 2 diabetes and cardiovascular disease in South Asians. Prim Care Diabetes 2011;5:45-56. CrossRef

8. American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care 2010;33(Suppl 1):S62-9. CrossRef

9. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med 1999;130:461-70. CrossRef

10. Chobanian AV, Bakris GL, Black HR, et al. Seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Hypertension 2003;42:1206-52. CrossRef

11. Gatineau M, Mathrani S. Obesity and ethnicity. Available from: http://www.noo.org.uk/uploads/doc/vid_9444_Obesity_and_ethnicity_270111.pdf. Accessed Jan 2011.

12. Gupta M, Singh N, Verma S. South Asians and cardiovascular risk: what clinicians should know. Circulation 2006;113:e924-9. CrossRef

13. Rees SD, Britten AC, Bellary S, et al. The promoter polymorphism -232C/G of the PCK1 gene is associated with type 2 diabetes in a UK-resident South Asian population. BMC Med Genet 2009;10:83. CrossRef

14. Samad S, Fatima J, Asma M. Prevalence of diabetes in Pakistan. Diabetes Res Clin Pract 2007;76:219-22. CrossRef

15. WHO STEPS Surveillance: Non Communicable Disease Risk Factors Survey. Kathmandu: Ministry of Health and Population, Government of Nepal, Society for Local Integrated Development Nepal (SOLID Nepal) and WHO; 2008.

16. Sharma D, Bkc M, Rajbhandari S, et al. Study of prevalence, awareness, and control of hypertension in a suburban area of Kathmandu, Nepal. Indian Heart J 2006;58:34-7.

Patent ductus arteriosus (PDA) is a common problem in preterm infants. Its occurrence is associated with prematurity and respiratory distress syndrome (RDS).1 2 A persistent left to right shunt in preterm neonates may be associated with neonatal morbidities, including bronchopulmonary dysplasia (BPD), intraventricular haemorrhage (IVH), and necrotising enterocolitis (NEC).3

Pharmacological closure of PDAs with indomethacin was first described in 1970s.4 Reported complications associated with the use of indomethacin included renal impairment,5 NEC, spontaneous intestinal perforation,6 and impaired cerebral blood flow.7 Ibuprofen, another cyclo-oxygenase inhibitor, has been investigated as an alternative to indomethacin for the same purpose. Published randomised controlled trials reported that ibuprofen was as efficacious as indomethacin for PDA closure, and some studies claimed that it had fewer adverse effects and gave rise to less renal impairment than indomethacin.8 9 10 11

The use of ibuprofen for closure of PDA has been increasing in clinical practice worldwide. In Hong Kong, indomethacin has been replaced by ibuprofen since 2010 due to interruption of the supply of indomethacin from the pharmaceutical company. Local data on its effectiveness and safety in clinical practice are very limited. A study comparing the use of ibuprofen versus indomethacin for this purpose could provide valuable data for clinicians regarding their use in clinical practice. At our unit, intravenous indomethacin had been used for treatment of PDA in preterm infants until April 2010. After that date, ibuprofen was used due to cessation of the supply of indomethacin from the pharmaceutical company supplying our hospital. We therefore set out to compare the two infant cohorts for treatment effectiveness and complications when used in our local setting.

This retrospective study was conducted in the neonatal intensive care unit (NICU) of Queen Elizabeth Hospital, a tertiary referral centre in Hong Kong with a level III neonatal intensive care service. The subjects in this study were all preterm infants admitted to the unit with their date of birth from 1 January 2008 to 31 December 2011 inclusive, and who had received at least one course of medical treatment for closure of a PDA with either indomethacin or ibuprofen. Due to the total switch from indomethacin to ibuprofen in clinical practice for this purpose in April 2010, we had information on two groups of infants—the indomethacin cohort (date of birth from 1 January 2008 to April 2010) and the ibuprofen cohort (date of birth from l April 2010 to 31 December 2011).

Preterm infants were defined as those who were born with less than 37 weeks of gestation. In our unit, preterm infants with clinical features suggestive of PDA, namely heart murmur, hypotension, hyperactive precordium, and increased ventilator settings were assessed by paediatric cardiologists. The diagnosis was then confirmed by echocardiography. Infants with a haemodynamically significant PDA were evaluated for medical closure with indomethacin/ibuprofen. Corresponding infants with features of heart failure, hypotension or who were ventilator-dependent were considered to have a haemodynamically significant PDA. Baseline assessments of these patients included platelet count, serum creatinine and electrolytes levels, urine output, and cranial ultrasound. Common contra-indications for the receipt of indomethacin/ibuprofen included thrombocytopenia, bleeding tendency, progressing IVH, NEC, and impaired renal function. Indomethacin was given at 0.1 mg/kg intravenously at 24-hour intervals for six doses or 0.2 mg/kg intravenously every 24 hours for three doses. Ibuprofen was given at 10 mg/kg, 5 mg/kg, and 5 mg/kg intravenously every 24 hours for a total of three doses. During the treatment courses, the infants were monitored for potential drug side-effects. Enteral feeding was withheld during the treatment course. Ductal closure was defined as persistent disappearance of the heart murmur; some of whom also had echocardiographic confirmation. Infants who failed the first course of medical treatment were re-evaluated and received a second course. Infants who failed two courses of medical treatment were considered for surgical ligation of the PDA in another tertiary referral centre in Hong Kong. Apart from the switch from indomethacin to ibuprofen in April 2010, the clinical practice for PDA management remained unchanged.

Eligible infants were identified by the Clinical Data Analysis and Reporting System, and their medical records were retrieved for data extraction. The neonatal demographic variables and baseline characteristics of both groups were collected and compared. The effectiveness of the drugs was primarily measured by (1) the failure rate of PDA closure after medical treatment, and (2) rate of recourse to surgical ligation. Secondary outcomes included all-cause mortality before discharge, BPD, adverse effects on renal function, gastro-intestinal complications (NEC and spontaneous intestinal perforation), and IVH. Occurrence of BPD was defined as (1) the use of supplement oxygen at 28 days of life or (2) the use of supplement oxygen at 36 weeks’ postmenstrual age. Adverse effects on renal function were inferred by the magnitude of any serum creatinine and/or urine output change. Necrotising enterocolitis was diagnosed and classified according to modified Bell’s staging.12 Intraventricular haemorrhage was classified according to the standard grading system.13

The two groups of infants receiving indomethacin or ibuprofen were compared using independent sample t tests for continuous normally distributed data, while the Wilcoxon rank-sum test was used for continuous non-normal data. Chi squared and Fisher’s exact tests were used as appropriate for categorical variables. The relative risks (RRs) of the outcome measures between the two groups were determined. The extent of change in urine output and serum creatinine level during the treatment course (within-subject effect) and the difference in change between the two groups (between-subject effect) were analysed by repeated measures analysis of variance. Potential confounding factors for medical closure of the PDA,8 including gender, gestational age, RDS grading, PDA ductal diameter, and day of starting treatment were evaluated by logistic regression. Significant factors were then entered into a multivariate logistic regression model to determine adjusted odds ratios. In all the analyses, a P value of less than 0.05 was considered significant. The statistical analysis was performed using the Statistical Package for the Social Sciences (Windows version 16.0; SPSS Inc, Chicago [IL], US). The sample size estimation was based on the primary outcome measure: the difference in proportion of infants with medical closure between the two groups. The sample size calculation for a moderate effect size of 0.3, power of 80%, and an alpha of 0.05 indicated that around 40 subjects were needed in each group.

This study was approved by the Research Ethics Committee, Kowloon Central Cluster, Hospital Authority.

In all, 96 infants had medical treatment for closure of a PDA during the study period; 52 (55%) received indomethacin only and 43 (45%) received ibuprofen only. One infant, who received both indomethacin and ibuprofen during the transitional period, was excluded. There were no significant differences in the demographic variables and baseline characteristics of the two groups (Table 1), except for a higher proportion with severe IVH (grades 3 and 4) in the indomethacin group (P=0.01). There was no significant difference between these groups with respect to the number of infants receiving one or two courses of treatment (Table 1).

Regarding the effectiveness of treatment, 20 (38%) of the infants in the indomethacin group and 18 (42%) in the ibuprofen group failed medical treatment after the first course; the RR of failure for the latter compared to indomethacin was 1.09, the 95% confidence interval (CI) being 0.69 to 1.69 (Table 2). Considering all courses of treatment with indomethacin or ibuprofen, 16 (31%) in the former group and 14 (33%) in the latter group failed medical treatment. Eleven (21%) infants in the indomethacin group and seven (16%) in the ibuprofen group underwent surgical ligation of the PDA. For the primary outcome measure, both groups were very comparable.

Factors with a potential to affect medical closure of PDA were evaluated. Among them, gestational age, RDS, and age at the start of medical treatment were shown to be significantly related to the rate of surgical ligation of PDA, with borderline significance for age at start of treatment in the univariate analysis (Table 3). When the above-mentioned significant factors were used in the multivariate analysis model, there was no significant difference between the two groups in terms of the rate of surgical ligation (adjusted odds ratio=0.94; 95% CI, 0.27-3.26; P=0.923).

Within the study cohort, two (4%) infants in the indomethacin group and five (12%) in the ibuprofen group died before being discharged, but this yielded no statistically significant difference in all-cause mortality.

The rates of BPD were also similar in both groups (P=0.615 for use of supplement oxygen at 28 days and P=0.560 for use of supplement oxygen at 36 weeks’ postmenstrual age). The mean duration of invasive ventilation for the indomethacin group, however, was significantly longer than that for ibuprofen group (mean ± standard deviation, 35 ± 35 vs 20 ± 25 days; P=0.045), while the mean duration of oxygen dependency was similar (P=0.694; Table 4).

Although not statistically significant, there was a higher rate of spontaneous intestinal perforation in the ibuprofen group (5% vs 0%, P=0.202), a higher rate of NEC (23% vs 12%, P=0.129), and NEC stage 2 or above (7% vs 2%, P=0.325) in the ibuprofen group (Table 4). On the other hand, on considering infants with NEC or spontaneous intestinal perforation together, there was a significantly higher rate in the ibuprofen than indomethacin group (P=0.043), and the same was true for gastro-intestinal bleeding (P=0.024).

Mean baseline serum creatinine concentrations and urine outputs were similar in the two groups (Table 1). Renal function related to the first course of treatment with indomethacin or ibuprofen was also studied. For within-subject effects, there were significant decreases in urine output (P<0.001) and increases in serum creatinine level (P=0.004) over time during treatment. For between-subjects effect, there was no significant difference in the changes of serum creatinine (P=0.829) and urine output (P=0.498) in the two groups, indicating that both drugs had a significant and comparable effect on renal function as measured by serum creatinine level and urine output (Fig).

A larger proportion of infants in the indomethacin group had severe IVH at baseline. However, in both groups the rates of progression of IVH after treatment were similar (P=0.644).

Our study compared the effectiveness and side-effects of intravenous indomethacin versus ibuprofen in treating PDA in preterm infants in two cohorts of Hong Kong patients. Our results demonstrate no significant difference in baseline characteristics between the two groups, thus justifying comparison of the cohorts. The two drugs appear to have similar effectiveness as measured by the rate of medical closure and surgical ligation rate of PDAs; such finding was also consistent with previous randomised controlled trials8 9 10 11and cohort studies.14 15 Even after potential confounding factors (discussed in the previous literature8) were controlled, the effectiveness of the two drugs did not differ significantly.

A higher all-cause mortality rate was observed in the ibuprofen group, although this did not reach statistical significance. The mortality case analysis was limited by the small number of deaths in each group; the power calculated was only 25%. Similar findings were reported by Katakam et al.15 When considering the individual cases, we observed that two infants in the ibuprofen group might have died of drug-related complications, namely spontaneous intestinal perforation and acute renal failure. Although one should not be biased by individual cases, these deaths illustrate the potential for fatal complications related to this drug.

We found no significant difference in the risk of BPD in the two groups; such result was consistent with that of a recent Cochrane review.16 By contrast another review by Jones et al17 concluded that intravenous ibuprofen may be associated with an increased risk of BPD when compared with intravenous indomethacin. These inconsistencies may be related to the definitions of BPD that were used. Our study considered BPD using the two most commonly used definitions (supplement oxygen use at 28 days and at 36 weeks' postmenstrual age, separately). Notably, similar rates of BPD were observed in the two groups for both definitions. By contrast, the duration of invasive ventilation was significantly longer in the indomethacin group. In this respect, a possible explanation and limitation of our study was that there may have been a gradual change in ventilation strategy over time, with a trend towards non-invasive ventilation.18

Regarding evaluation of possible gastrointestinal complications, two conditions (NEC and spontaneous intestinal perforation) have been described. Both are believed to be associated with impaired mesenteric blood flow due to a PDA as well as the use of cyclo-oxygenase inhibitors, though some recent studies have reported on the difference in clinical presentations and histological findings between these two entities.19 20 We observed a statistically higher rate of intestinal complications (NEC or spontaneous intestinal perforation) in the ibuprofen group (P=0.043). In contrast, the latest Cochrane review16 reported less NEC in the ibuprofen group (RR=0.68; 95% CI, 0.47-0.99). The management practice of preterm infants in our unit, including the feeding regimen, remained unchanged during the study period. Thus, this particular inconsistency could not be attributed to any known factors. Kushnir and Pinheiro14 studied 350 infants and also reported a higher rate of NEC in ibuprofen than indomethacin users (8% vs 4%; P=0.08). Rao et al19 studied 102 infants with PDA treated with ibuprofen, and reported a 9% rate of spontaneous intestinal perforation and 6% rate of NEC; such figures were comparable to those in our ibuprofen cohort. These findings suggest that compared with preterm infants treated with indomethacin, intestinal complications appear to be more common in those receiving ibuprofen.

We found that indomethacin and ibuprofen had a similar effect on renal function, though previous literature 8 9 15 16 17indicated that ibuprofen had less effect on renal blood flow and renal function. This inconsistency could be related to differences in how measurement of renal function was carried out. We evaluated the change in serum creatinine and urine output during the course of treatment. The change in these parameters, rather than the absolute values, might be better parameters to assess due to variations in serum creatinine with gestational age and the age of the infants.21 Another problem was the timing of measurements. Akima et al22 evaluated the renal effects of indomethacin and reported a significant increase in serum creatinine level on day 2 and day 7 of treatment when compared with the controls. Due to differences in the duration of treatment courses with the two drugs, the best time to carry out comparisons remains unclear. Moreover, as observed in one of our infants given ibuprofen who also developed acute renal failure 2 days after the completion of second course, there could be delayed and cumulative effects on renal function with repeat treatment courses. This was also shown by Kushnir and Pinheiro,14 whereby indomethacin had a more prominent effect on renal function during the first course while both drugs led to equal adversity at the second and third courses. However, the retrospective design of our study was a limitation as some data (especially on day 4 and later) in the ibuprofen group were influenced by the course lasting only 3 days, whilst data on the repeat courses of treatment were less complete. Hence, our study evaluated the first 4 days of the first course of treatment, and evaluation of repeat courses was excluded. With regard to the significant renal effects of ibuprofen and indomethacin noted in our study, we recommend close monitoring of renal function when either drug is used. Special cautions may be necessary for repeat courses of treatment.

Till now, published studies on the efficacy and safety of ibuprofen versus indomethacin were mainly randomised trials. The subjects in randomised trials were selected using inclusion and exclusion criteria that may be less representative of the whole spectrum of infants in clinical practice. For instance, randomised trials by Van Overmeire et al8 and Lago et al9 only studied infants with PDA treatment given in the first 2 to 4 days of life and RDS was an inclusion criterion. Our study included all infants that were treated within the study period, maximising the representativeness of the sample. Being a retrospective study to investigate the effectiveness and complications related to drug therapy in a clinical setting, the allocation of treatment was not randomised or blinded. However, selection bias was minimised as the drug treatment each infant received was only determined by the month and year they were admitted to the neonatal unit. On the other hand, being a study from two contiguous time periods, there may have been minor modifications of clinical practice despite that both infant cohorts being managed by the same group of clinicians and there being no change in departmental guidelines for management of PDAs.

Our study shared the limitations of most previous studies. Our sample size estimation was based on the primary outcome (the rate of successful medical closure). As the sample size was limited by the number of eligible infants within the study period, the effect size adopted in the sample size estimation was 0.3, which was moderate compared to other similar studies. Moreover, with respect to adverse outcome evaluation, infant numbers with positive findings were small, which affected the precision of our analyses. Another limitation was that two regimens of the indomethacin were used in our hospital: 0.1 mg/kg/dose every 24 hours for six doses (prolonged course) and 0.2 mg/kg/dose every 24 hours for three doses (short course). Fortunately, this heterogeneity within the group was small, as the majority of infants received the prolonged course (46 out of 52). Moreover, previous studies comparing these two regimens showed that their efficacy did not differ significantly.23 24 As for the generalisability of our study, variations in management of symptomatic PDA do exist between centres,25 26 and there is no consensus approach. Our practice, for trial of a second course of indomethacin or ibuprofen before considering surgical ligation, entailed intense monitoring for adverse effects, which was consistent with common practice.23 Thus, our study could provide useful information for other NICUs to consider for the management of PDA in preterm infants.

In clinical practice, intravenous ibuprofen is as effective as indomethacin for the medical closure of PDAs in premature infants. However, owing to the higher rates of intestinal complications after ibuprofen therapy, we conclude that it may not have fewer adverse effects than indomethacin. Neonatologists are therefore advised to cautiously monitor for possible side-effects in preterm infants receiving either indomethacin or ibuprofen for the treatment of PDAs.

No conflicts of interests were declared by authors.

1. Kitterman JA, Edmunds LH Jr, Gregory GA, Heymann MA, Tooley WH, Rudolph AM. Patent ductus arteriosus in premature infants — incidence, relation to pulmonary disease and management. N Engl J Med 1972;287:473-7. CrossRef

2. Thibeault DW, Emmanouilides GC, Nelson RJ, Lachman RS, Oh W. Patent ductus arteriosus complicating the respiratory distress syndrome in preterm infants. J Pediatr 1975;86:120-6. CrossRef

3. Clyman RI, Chorne N. Patent ductus arteriosus: evidence for and against treatment. J Pediatr 2007;150:216-9. CrossRef

4. Friedman WF, Hirschklau MJ, Printz MP, Pitlick PT, Kirkpartrick SE. Pharmacologic closure of patent ductus arteriosus in the premature infant. N Engl J Med 1976;295:526-9. CrossRef

5. Seyberth HW, Rascher W, Hackenthal R, Wille L. Effect of prolonged indomethacin therapy on renal function and selected vasoactive hormones in very low birth weight infants with symptomatic patent ductus arteriosus. J Pediatr 1983;103:979-84. CrossRef

6. Fujii AM, Brown E, Mirochnick M, O'Brien S, Kaufman G. Neonatal necrotizing enterocolitis with intestinal perforation in extremely premature infants receiving early indomethacin treatment for patent ductus arteriosus. J Perinatol 2002;22:535-40. CrossRef

7. Edwards AD, Wyatt JS, Richardson C, et al. Effects of indomethacin on cerebral haemodynamics in very preterm infants. Lancet 1990;335:1491-5. CrossRef

8. Van Overmeire B, Smets K, Lecoutere D, et al. A comparison of ibuprofen and indomethacin for closure of patent ductus arteriosus. N Engl J Med 2000;343:674-81. CrossRef

9. Lago P, Bettiol T, Salvadori S, et al. Safety and efficacy of ibuprofen versus indomethacin in preterm infants treated for patent ductus arteriosus: a randomised controlled trial. Eur J Pediatr 2002;161:202-7.  CrossRef

10. Mosca F, Bray M, Lattanzio M, Fumagalli M, Tosetto C. Comparative evaluation of the effects of indomethacin and ibuprofen on cerebral perfusion and oxygenation in preterm infants with patent ductus arteriosus. J Pediatr 1997;131:549-54. CrossRef

11. Su BH, Lin HC, Chiu HY, Hsieh HY, Chen HH, Tsai YC. Comparison of ibuprofen and indomethacin for early-targeted treatment of patent ductus arteriosus in extremely premature infants: a randomized controlled trial. Arch Dis Child Fetal Neonatal Ed 2008;93:F94-9. CrossRef

12. Kliegman RM, Walsh MC. Neonatal necrotizing enterocolitis: pathogenesis, classification, and spectrum of illness. Curr Probl Pediatr 1987;17:213-88. CrossRef

13. Papile LA, Burstein J, Burstein R, Koffler H. Incidence and evolution of subependymal and intraventricular hemorrhage: a study of infants with birth weights less than 1,500 gm. J Pediatr 1978;92:529-34. CrossRef

14. Kushnir A, Pinheiro JM. Comparison of renal effects of ibuprofen versus indomethacin during treatment of patent ductus arteriosus in contiguous historical cohorts. BMC Clin Pharmacol 2011;11:8. CrossRef

15. Katakam LI, Cotton CM, Goldberg RN, Dang CN, Smith PB. Safety and effectiveness of indomethacin versus ibuprofen for treatment of the patent ductus arteriosus. Am J Perinatol 2010;27:425-9. CrossRef

16. Ohlsson A, Walia R, Shah SS. Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants. Cochrane Database Syst Rev 2010;(4):CD003481.

17. Jones LJ, Craven PD, Attia J, Thakkinstian A, Wright I. Network meta-analysis of indomethacin versus ibuprofen versus placebo for PDA in preterm infants. Arch Dis Child Fetal Neonatal Ed 2011;96:F45-52. CrossRef

18. Ramanathan R, Sardesai S. Lung protective ventilatory strategies in very low birth weight infants. J Perinatol 2008;28 Suppl 1:S41-6. CrossRef

19. Rao R, Bryowsky K, Mao J, Bunton D, McPherson C, Mathur A. Gastrointestinal complications associated with ibuprofen therapy for patent ductus arteriosus. J Perinatol 2011;31:465-70. CrossRef

20. Attridge JT, Clark R, Walker MW, Gordon PV. New insights into spontaneous intestinal perforation using a national data set: (2) two populations of patients with perforations. J Perinatol 2006;26:185-8. CrossRef

21. Bueva A, Guignard JP. Renal function in preterm neonates. Pediatr Res 1994;36:572-7. CrossRef

22. Akima S, Kent A, Reynolds GJ, Gallagher M, Falk MC. Indomethacin and renal impairment in neonates. Pediatr Nephrol 2004;19:490-3. CrossRef

23. Tammela O, Ojala R, Iivainen T, et al. Short versus prolonged indomethacin therapy for patent ductus arteriosus in preterm infants. J Pediatr 1999;134:552-7. CrossRef

24. Siu KL, Wan KM, Law CW, Lee WH. Indomethacin regimes for patent ductus arteriosus in premature newborn. HK J Paediatr (New Series) 1997;2:180-1.

25. Amin SB, Handley C, Carter-Pokras O. Indomethacin use for the management of patent ductus arteriosus in preterms: a web-based survey of practice attitudes among neonatal fellowship program directors in the United States. Pediatr Cardiol 2007;28:193-200. CrossRef

26. Brissaud O, Guichoux J. Patent ductus arteriosus in the preterm infant: a survey of clinical practices in French neonatal intensive care units. Pediatr Cardiol 2011;32:607-14. CrossRef

Significant unprotected left main coronary artery (ULMCA) disease occurs in 5% to 7% of patients undergoing coronary angiography.1 Coronary artery bypass graft (CABG) surgery has been the standard of care for the treatment of ULMCA disease, and percutaneous coronary intervention (PCI) is reserved for patients who are poor surgical candidates.2 Recently, the use of drug-eluting stents (DES), together with advance in PCI technology, has improved the outcomes of patients undergoing PCI for ULMCA disease. The latest guidelines assign ULMCA PCI a class IIa indication which may be considered in patients who are at low risk for procedural complications and at increased risk of adverse surgical outcomes.3

Because of the risk of restenosis, it is not encouraged to use bare-metal stents (BMS) in ULMCA disease. The situation in Hong Kong is special in this regard. The public health care system (The Samaritan Fund) of Hong Kong does not cover the cost of using DES in ULMCA disease. Hence, patients with financial difficulty and who refuse to receive CABG can only undergo PCI with BMS implantation. Moreover, like other Asian countries, patients in Hong Kong are reluctant to have CABG, leaving them with the option of using BMS or medical treatment only. Because of this restraint, the proportion of patients with ULMCA disease in Hong Kong who are treated with BMS probably exceeds that in other developed countries.

The present study aimed to evaluate the outcomes of patients with ULMCA stenosis who were treated with PCI in Hong Kong.

This was a single-centre retrospective study performed to determine the outcomes of patients who had undergone ULMCA PCI. Between January 2008 and September 2011, 111 patients with ULMCA disease (defined as >50% stenosis) received PCI with either DES or BMS implantation in the United Christian Hospital, Hong Kong. The cohort included unselected consecutive patients who presented with stable angina, acute coronary syndrome, or cardiogenic shock. Therefore, PCI could be performed in an elective or emergency setting (ie an all-comers basis). Moreover, there was no on-site surgical support in our centre.

The decision of performing PCI instead of CABG surgery was based on coronary anatomy, haemodynamic conditions, surgical risks, and patients’ preference. Both interventional cardiologists and cardiac surgeons were involved in making the decision.

Unprotected left main coronary artery PCI was performed using standard techniques. Heparin 70 to 100 units per kg was administered before PCI. Intra-aortic balloon pump counterpulsation, intravascular ultrasound (IVUS) or glycoprotein IIb/IIIa inhibitors was used at the discretion of the operators. All patients were pre-treated with 80 to 160 mg aspirin and a loading dose of 300 to 600 mg clopidogrel or 75 mg maintenance dose of clopidogrel at least 7 days before the procedure. After PCI, aspirin 80 to 160 mg daily and clopidogrel 75 mg daily, for 1 month after BMS and 1 year after DES implantation, were prescribed. For ostial and shaft left main stenosis, single stent placement was preferred. Patients with bifurcation stenosis underwent one of the four types of bifurcation stenting techniques (T-stenting, T-stenting and small protrusion technique, Culotte technique, or Crush technique) at the operators’ discretion. Routine surveillance angiography was arranged for all patients 6 to 9 months after the index procedure, except in patients who refused, or with high risk for coronary angiogram. Baseline demographic, procedural, angiographic, and clinical outcome data were collected.

Unprotected left main coronary artery stenosis was defined as >50% stenosis without any patent graft to the left anterior descending artery or left circumflex artery. Procedure was defined as successful if revascularisation was achieved in the target lesion with <30% residual stenosis in angiography and patient was discharged from hospital without any of these events: death, Q-wave myocardial infarction (MI), stroke, and target lesion revascularisation (TLR).

Follow-up was completed in June 2012. End-points were restenosis and major adverse cardiac and cerebrovascular events (MACCE) including cardiac death, non-fatal MI, stroke, and TLR.

Restenosis was defined as >50% luminal narrowing at the left main segment (stent and 5 mm proximal and distal) which was demonstrated at the follow-up angiography, regardless of patient symptoms.

Death was classified as cardiac or non-cardiac. Deaths that could not be classified were considered cardiac. Cardiac death was defined as death from any cardiac cause (eg MI, heart failure, or arrhythmia) or sudden unexplained death without an explanation. Non–Q-wave MI was defined as elevation of total creatine kinase 2 times above the upper normal limit in the absence of pathological Q wave. Target lesion revascularisation was defined as any revascularisation performed on the treated left main segment. Chronic kidney disease was documented if the serum creatinine level was >200 µmol/L or was put on renal replacement therapy. Stent thrombosis was defined as definite and probable according to the Academic Research Consortium.4

Categorical variables reported as percentages and comparisons between groups were based on the Chi squared test or Fisher’s exact test. Continuous variables were reported as mean ± standard deviation, and differences were assessed with the independent sample t test or Mann-Whitney test.

Cumulative event curves were calculated by the Kaplan-Meier method and compared by the log-rank test. A P value of <0.05 was considered statistically significant. Statistical analyses were performed with the use of the Statistical Package for the Social Sciences (Windows version 15.0; SPSS Inc, Chicago [IL], US).

Baseline clinical, and angiographic and procedural characteristics of the 111 patients are summarised in Table 1and Table 2, respectively.

Overall, 86 (77.5%) patients were treated with DES, and 25 (22.5%) received BMS. The two groups shared similar clinical and angiographic characteristics. More than 90% of patients had left ventricular ejection fraction of ≥35%. The majority of patients had distal left main disease (81.4% in DES group and 72.0% in BMS group). Only a minority of patients (5.4%) had isolated left main disease, whereas 72.9% had left main and at least two-vessel disease. A high rate of IVUS use was observed in the cohort (84.7%). Final kissing balloon dilatation was performed in >50% of the patients and in all patients with two-stent approach. Other adjuvant PCI devices such as rotational atherectomy were rarely required in this cohort.

Of the 86 patients who received DES at the left main segment, 24 (27.9%) received first-generation DES, 56 (65.1%) received second-generation DES, and six (7.0%) received both types.

Procedural success was achieved in 109/111 (98.2%) cases. There was one death (0.9%) and one stroke (0.9%) but there was no Q-wave MI, stent thrombosis, or urgent repeat revascularisation events during hospitalisation (Table 3).

The mean duration of clinical follow-up was 26.1 ± 12.6 months. Table 4 depicts the incidence of adverse outcomes in all patients at the end of follow-up. There was no significant difference between the DES and BMS groups in the cumulative incidences of cardiac death (5.8% for DES vs 16.0% for BMS; P=0.191) or non-fatal MI (3.5% vs 8.0%; P=0.262). Compared with BMS, use of DES was associated with significantly lower risks of TLR (9.3% vs 32.0%; P=0.001) and MACCE (19.8% vs 44.0%; P=0.004) [Fig]. Target lesion revascularisation was ischaemia-driven in 4/16 (25%) patients; in the remaining 12/16 (75%) patients, TLR was driven by restenosis identified at surveillance angiography after the index procedure. Therefore, the crude rate of ischaemia-driven TLR was only 4/111 (3.6%) in the overall cohort. The mean timing of TLR was 7.6 ± 4.3 months (range, 2-16 months) after the index procedure.

Of 111 cases, 93 (83.8%) underwent routine surveillance angiography 6 to 9 months after PCI; binary restenosis occurred in 22/111 (20%) cases. Restenosis occurred predominantly in patients with distal left main coronary artery disease (19/22 [86%]); and more than half of them (12/22 [55%]) had isolated focal restenosis involving the ostium of the left circumflex artery only. Restenosis occurred less frequently with DES than with BMS (12/86 [14.0%] vs 10/25 [40.0%]; P=0.004).

For stent thrombosis, the event rate was extremely low across the whole cohort. One patient receiving BMS implantation developed subacute stent thrombosis after hospital discharge (which resulted in sudden cardiac death). There was no stent thrombosis of any forms in the DES group.

The principal findings of the present study were: (1) performing PCI for ULMCA disease was safe and feasible in selected patients with high procedural success rate (98.2%); (2) after an intermediate-term follow-up of 26.1 months, the incidence of MACCE in patients receiving DES implantation was similar to that reported in recent major international clinical trials including the SYNTAX trial5; (3) compared with BMS, the use of DES was associated with a lower risk of restenosis and repeat revascularisation without an increased risk of death or MI.

Historically, CABG has been regarded as the gold standard of treatment for ULMCA disease. Clinical outcomes after PCI for ULMCA stenosis have been shown to vary widely, according to patients’ clinical and angiographic features.6 7 The high procedural success rate in our study further confirms the technical feasibility of treating ULMCA lesions with the current PCI techniques in the absence of on-site surgical support.

Promising results were reported from randomised trials comparing first-generation DES versus CABG.5 8 9 In the SYNTAX trial,5 patients were stratified according to the presence of ULMCA disease and randomised to CABG (n=348) or PCI with paclitaxel-eluting stents (n=357). In the ULMCA subgroups, MACCE at 12 months was comparable between patients treated with PCI and CABG. Moreover, although the rate of repeat revascularisation among patients with ULMCA disease was significantly higher in the PCI subgroup, this result was offset by a significantly higher rate of stroke in the CABG subgroup.

The SYNTAX trial5 included patients with heterogeneous angiographic characteristics in the left main subgroup (13% with isolated left main coronary artery disease, 20% with left main plus single-vessel disease, 31% with two-vessel disease, and 37% with triple-vessel disease). Although calculation of the SYNTAX score was not incorporated in routine clinical practice at the time of our study, our cohort demonstrated similar heterogeneity and complexity (Table 2).

We report an intermediate-term outcome (mean follow-up of approximately 26 months) for patients with ULMCA PCI, and our results were comparable with those of the SYNTAX trial.5 At 2 years, the SYNTAX trial5 reported a MACCE rate of 22.9% in the left main subgroup (including death from any causes, MI, stroke, or repeat revascularisation), which was comparable with the incidence of 19.8% reported in our study.

The incidence of TLR in the subgroup of DES in our registry (9.3%) might be lower than that reported in the SYNTAX trial5 at 2 years (any revascularisation, 17.3%) and it might be due to inclusion of second-generation DES in two thirds of the patients treated with DES in our registry. The higher rate of IVUS use for optimisation (approximately 90% of cases using DES in our cohort) might also be another reason. One of the main limitations of the SYNTAX trial was thought to be the lack of IVUS use for ULMCA disease in the PCI group. Clinical trials10 have shown that patients whose coronary interventions are guided by IVUS have larger post-procedure stent areas and significant reductions in TLR than those undergoing angiography-guided PCI only. Registry data have also shown a trend towards reduced mortality in IVUS-guided ULMCA PCI.11

It is worth considering that SYNTAX did not have an ‘all-comers’ design, where patients with acute coronary syndrome and cardiogenic shock were excluded. Our registry did have an ‘all-comers’ design, by including patients presenting with stable angina, acute coronary syndrome, ST-elevation and non–ST elevation MI, as well as cardiogenic shock. This might reflect a more ‘real-world’ situation in daily clinical practice. Despite the inclusion of patients with higher clinical risk, the incidence of events remained low in our study during the index hospital admission and upon medium-term follow-up.

In the BMS subgroup, we reported a high incidence of restenosis (40%) and TLR (32%). To date, no randomised controlled trials have been performed using BMS in ULMCA PCI. The longest follow-up available in the literature was from the ASAN-MAIN (ASAN Medical Center–Left MAIN Revascularization) Registry (n=350: BMS, n=100; CABG, n=250),12 which also reported a high rate of TLR (24.9%) after long-term follow-up. Although the incidence of restenosis and TLR might be over-represented due to the use of routine surveillance angiography in our study, the results suggest that the use of BMS was not favoured.

As mentioned, the situation in Hong Kong is unique in that the public health care system does not cover the cost of using DES in ULMCA disease. Patients with financial difficulty can only choose PCI with BMS or CABG. Because of this restraint, the proportion of patients with ULMCA disease in Hong Kong treated with BMS probably exceeds that in other developed countries. In our opinion, a review of this health care policy is necessary.

In our cohort, the rate of cardiac deaths in the BMS group was relatively high (16.0% in BMS vs 5.8% in DES). While this could be a finding by chance, it could be attributed to a multitude of reasons. Compared with the DES group, a higher proportion of patients presented with acute coronary syndrome including cardiogenic shock in the BMS group (Table 1). Moreover, there was a higher proportion of patients with chronic renal failure or prior stroke in the BMS group (Table 1). Such differences might explain the relatively high cardiac mortality rates in the BMS group. Another postulation is that patients who received BMS implantation may have come from a lower socio-economic class, which might have an impact on their health status and outcome.

The role of routine surveillance angiography remains unclear and controversial. Repeat angiography is suggested because patients with left main restenosis are considered to be at high risk for adverse events. However, angiography is unable to predict when a patient might be prone to stent thrombosis, and angiography might be associated with a non-negligible risk in patients who have undergone left main stenting.13 Therefore, the 2009 focused update does not recommend routine angiographic follow-up after ULMCA stenting.14 Our result is in line with the guideline as the angiographic restenosis rate in the DES group was low. This would have been even lower had a clinically driven approach been used. Given the low event rate in our cohort, we also recommend that routine surveillance angiography is not necessary and patients can be followed up clinically.

An interesting point is that the risk of stent thrombosis was extremely low (<1%) given the standard prescription of 1-year dual antiplatelet therapy with aspirin and clopidogrel in this group of high-risk patients with multiple complex stenting. No laboratory or genetic assessment was performed on the degree of platelet function inhibition.

The present study had several limitations. Firstly, it was a single-centre non-randomised retrospective study, which might have significantly affected the results due to unmeasured confounders, procedure bias, or detection bias. Secondly, angiographic results were based on visual angiographic or IVUS assessment and a standardised core laboratory anatomical examination was not performed. Thirdly, incomplete angiographic follow-up might underestimate the incidence of restenosis. Finally, this study included high-risk patients with complex coronary anatomy who underwent PCI (including patients who refused bypass surgery); these patients were prone to poor clinical outcomes. Therefore, these results might not be generalised to all populations with ULMCA stenosis, especially those with low-to-intermediate SYNTAX score.

These are the largest available data on ULMCA PCI in Hong Kong. Performing PCI for ULMCA disease was safe and feasible in selected patients with high procedural success. Despite the inclusion of high-risk patients, the incidence of MACCE after intermediate-term follow-up in patients receiving DES implantation was similar to that reported in major clinical trials. Compared with BMS, DES was associated with a reduced need for repeat revascularisation without increasing the risk of death or MI for patients with ULMCA disease. Our result suggest that BMS should not be encouraged due to the high incidence of restenosis and TLR.

The authors report no financial relationships or conflicts of interest regarding the content herein.

The authors wish to thank Dr CY Mui and Dr TK Lau for their assistance in data collection.

1. DeMots H, Rösch J, McAnulty JH, Rahimtoola SH. Left main coronary artery disease. Cardiovasc Clin 1977;8:201-11.

2. Eagle KA, Guyton RA, Davidoff R, et al. ACC/AHA 2004 guideline update for coronary artery bypass graft surgery: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee to Update the 1999 Guidelines for Coronary Artery Bypass Graft Surgery). Circulation 2004;110:1168-76. CrossRef

3. Levine GN, Bates ER, Blankenship JC, et al. 2011 ACCF/AHA/ SCAI guideline for percutaneous coronary intervention. A report of the American College of Cardiology Foundation/ American Heart Association Task Force on Practice Guidelines and the Society for Cardiovascular Angiography and Interventions. J Am Coll Cardiol 2011;58:e44-122. CrossRef

4. Cutlip DE, Windecker S, Mehran R, et al. Clinical end points in coronary stent trials: a case for standardized definitions. Circulation 2007;115:2344-51. CrossRef

5. Serruys PW, Morice MC, Kappetein AP, et al. Percutaneous coronary intervention versus coronary-artery bypass grafting for severe coronary artery disease. N Engl J Med 2009;360:961-72. CrossRef

6. Park SJ, Kim YH, Lee BK, et al. Sirolimus-eluting stent implantation for unprotected left main coronary artery stenosis: comparison with bare metal stent implantation. J Am Coll Cardiol 2005;45:351-6. CrossRef

7. Price MJ, Cristea E, Sawhney N, et al. Serial angiographic follow-up of sirolimus-eluting stents for unprotected left main coronary artery revascularization. J Am Coll Cardiol 2006;47:871-7. CrossRef

8. Park SJ, Kim YH, Park DW, et al. Randomized trial of stents versus bypass surgery for left main coronary artery disease. N Engl J Med 2011;364:1718-27. CrossRef

9. Boudriot E, Thiele H, Walther T, et al. Randomized comparison of percutaneous coronary intervention with sirolimus-eluting stents versus coronary artery bypass grafting in unprotected left main stem stenosis. J Am Coll Cardiol 2011;57:538-45. CrossRef

10. Hong MK, Mintz GS, Lee CW, et al. Intravascular ultrasound predictors of angiographic restenosis after sirolimus-eluting stent implantation. Eur Heart J 2006;27:1305-10. CrossRef

11. Park SJ, Kim YH, Park DW, et al. Impact of intravascular ultrasound guidance on long-term mortality in stenting for unprotected left main coronary artery stenosis. Circ interventions 2009;2:167-77.

12. Park DW, Kim YH, Yun SC, et al. Long-term outcomes after stenting versus coronary artery bypass grafting for unprotected left main coronary artery disease. J Am Coll Cardiol 2010;56:1366-75. CrossRef

13. Lee MS, Kapoor N, Jamal F, et al. Comparison of coronary artery bypass surgery with percutaneous coronary intervention with drug-eluting stents for unprotected left main coronary artery disease. J Am Coll Cardiol 2006;47:864-70. CrossRef

14. Kushner FG, Hand M, Smith SC Jr, et al. 2009 focused updates: ACC/AHA guidelines for the management of patients with ST-elevation myocardial infarction (updating the 2004 guideline and 2007 focused update) and ACC/ AHA/SCAI guidelines on percutaneous coronary intervention (updating the 2005 guideline and 2007 focused update). J Am Coll Cardiol 2009;54:2205-41. CrossRef