Customer advice on sports injuries from Worcester Physiotherapy Clinic Cortisone is a type of steroid that is produced naturally by a gland in your body called the adrenal gland. Cortisone is released from the adrenal gland when your body is under stress. Natural cortisone is released into the blood stream and is relatively short-acting. Injectable cortisone is a synthetic (man made) drug a
Pulsed-dye laser treatment for inflammatory acne vulgaris:randomised controlled trial E D Seaton, A Charakida, P E Mouser, I Grace, R M Clement, A C Chu IntroductionAcne vulgaris is a common disease that has been Background Low-fluence (low irradiation energy density) associated with social isolation, employment difficulties, pulsed-dye lasers (PDLs) have been used for atrophic acne depression, and suicide.1–4 The many treatments that are scarring, and anecdotal experience suggests that long-term available indicate the dissatisfaction of patients and improvements in inflammatory acne can be seen after one doctors with available therapies and difficulties in PDL treatment. Our aim was to compare the efficacy and management of this disease. New, effective, and well tolerability of such PDL treatment with sham treatment in patients with facial inflammatory acne in a double-blind, Early inflammatory acne lesions are characterised by the infiltration of the pilosebaceous duct with CD4+ T-helper-1 cells that are reactive to Propionibacterium Methods We recruited 41 adults with mild-to-moderate facial acnes, a common cutaneous commensal.5,6 Colonisation inflammatory acne. We randomly assigned patients to PDL of individuals with this bacterium is closely associated (n=31) or sham treatment (n=10). Treatment was given with the development of inflammatory acne, and the at baseline and patients were seen after 2, 4, 8, and development of antibiotic resistance of P acnes is 12 weeks. Assessors and participants were unaware of associated with treatment failure.7,8 P acnes is a porphyrin- treatment allocations. Primary outcome measures were containing organism that is killed by exposure to specific acne severity after 12 weeks and adverse events at any wavelengths of light.9,10 The photosensitivity of the time. Secondary measures were change in lesion counts bacterium accounts for the improvement noticed by most after 12 weeks and change in acne severity with time.
individuals with acne after exposure to sunlight, and has encouraged the development of artificial visible lightsources as treatment for this disease.11,12 Findings After 12 weeks, acne severity (measured by Leeds Lasers differ from non-laser light sources in that they revised grading system) was reduced from 3·8 (SD 1·5) to emit minimally divergent, coherent light that can be 1·9 (1·5) in the PDL group and 3·6 (1·8) to 3·5 (1·9) in the focused to a small area of tissue to provide very high sham group (p=0·007). Treatment was well tolerated. Total irradiances. Pulsed-dye lasers (PDLs) emit visible light lesion counts fell by 53% (IQR 19 to 64) in PDL patients and that is mainly absorbed by oxyhaemaglobin, so high 9% (–16 to 38) in controls (p=0·023), and inflammatory irradiation energy densities (fluences) are used to treat lesion counts reduced by 49% (30 to 75) in PDL patients vascular lesions such as port wine stains.13 Whereas high and 10% (–8 to 49) in controls (p=0·024). The most rapid fluences ablate small blood vessels and cause purpura, improvements were seen in the first 4 weeks after lower non-ablative fluences do not. Low fluences can, however, stimulate cutaneous procollagen production,secondary to a non-lethal heating of dermal perivascular Interpretation PDL therapy improves inflammatory facial tissues that is postulated to alter local cellular acne 12 weeks after one treatment with no serious adverse metabolism.14 Non-ablative PDLs are increasingly used in cosmetic practice to improve the appearance of finewrinkles and are effective in the treatment of atrophic Experience in several clinics suggests that a proportion of patients receiving low-fluence PDL treatment havecoincidental striking and longstanding improvements ininflammatory acne after a sole treatment of the face(unpublished). We aimed to examine the efficacy andtolerability of a single low-fluence PDL treatment inpatients with facial inflammatory acne. Department of Dermatology, Hammersmith Hospital Campus, Faculty of Medicine, Imperial College, London W12 0NN, UK Individuals were recruited through a public request for (E D Seaton MRCP, A Charakida MBBS, P E Mouser BSc, A C Chu FRCP); participants or because of referral to the dermatology Department of Primary Care and Population Health Sciences, clinic. Recruitment took place between Nov 13, 2001, Faculty of Medicine, Charing Cross Campus, Imperial College, and April 26, 2002, so that confounding effects of London (I Grace MSc); and School of Engineering, University of summer sunlight were avoided. Eligible patients were aged between 18 and 45 years with mild-to-moderate facial inflammatory acne defined as the presence of at least ten acne papules or pustules between the brow and THE LANCET • Vol 362 • October 25, 2003 • www.thelancet.com grading method in many clinical trialsof acne treatment.18 The investigatorshave had longstanding experience ofboth the use of this technique and of patients as an additional assessment.
Lesion counts were recorded for the assessed, every laser-allocated patientreceived treatment in which adifferent fluence was used on each side of the midline. Patents wererandomly allocated to receive 1·5 J/cm2 on one side of the midlineand 3·0 J/cm2 on the other. We used a laser spot diameter of 5 mm, andpulse duration of 350 µs (Nlite system, EUPhotonics, Swansea,Wales, UK). Patients’ whole faces jawline and an acne severity score of between 2 and 7 on were treated in about 15 min by moving the laser Washout periods for previous treatments were 4 weeks Controls were treated with a disconnected laser for oral antibiotics, 12 weeks for cyproterone acetate- handpiece that was moved across the face in an identical containing contraceptives, 52 weeks for oral isotretinoin, manner to that for the PDL group. All patients wore and 2 weeks for topical treatments. Acne treatments were opaque goggles during treatment to protect their eyes and not allowed during the study. The local ethics committee to ensure that they were unaware of the therapy they approved our protocol and all patients gave written received. Treatment was given in a locked room with no The primary endpoints of the study were change in acne severity after 12 weeks based on the Leeds revised At recruitment, patients were randomised to either laser grading system and adverse events at any time.
or a sham treatment by a computer-generated sequence.
Secondary endpoints were changes in total, inflammatory Allocations were contained in opaque, sequentially- and non-inflammatory lesion counts by the end of the numbered, sealed envelopes and were concealed from trial, and changes in acne severity with time. We also assessors and patients throughout the study and revealed assessed the proportion of patients achieving a reduction only to the investigator (EDS, AC, or ACC) who was of 1 or 2 points in acne grade or a 50% reduction in total assigned to treat the patient. Investigators were not acne lesion count by 12 weeks. Subgroup analysis of included in preliminary or post-treatment assessments of total, inflammatory, and non-inflammatory lesion counts on each side of the midline was done in laser-treated Patients received a single treatment at baseline and patients to allow assessment of the effect of different laser were reviewed after 2, 4, 8, and 12 weeks. For every patient, one trained investigator (EDS or ACC) recordeddemographic details and did clinical assessments with acne grading, total lesion counts (inflammatory and non- Data conforming to a normal distribution were analysed inflammatory lesions), inflammatory lesion counts with two-sample t tests. We analysed non-normally (papules and pustules), and non-inflammatory lesion counts (open and closed comedones).17 The Leeds independent groups and Wilcoxon matched pairs signed revised acne grading system is a rapid and reproducible rank test for paired data (half face comparisons).
means of recording inflammatory acne by matching acne Changes from baseline are reported in absolute numbers severity with validated photographs of acne patients and and percentages, with statistical analyses done for assigning a numerical score between 1 and 12. Patients absolute values. Analysis of proportional data was done with no active inflammatory acne were assigned a score of with Fisher’s exact test. For the primary clinical outcome THE LANCET • Vol 362 • October 25, 2003 • www.thelancet.com differences in time, treatment group allocation, and a combination of these two factors. Analyses of allsecondary endpoints were per protocol. We used an uneven allocation ratio of 3 to 1 to facilitate assessment of the safety of this previously unreported treatment and to encourage recruitment. This design improves the probability of identifying infrequent adverse events although, inevitably, reduces the power of the study to detect differences in efficacy between groups, by an amount that is equivalent to excluding a quarter of patients.20 Our data should allow calculation of sample The sponsors of the study had no role in study design, data collection, data analysis, data interpretation, or Figure 1 shows the trial profile. 26 laser-allocated patients and nine controls had volunteered for the trialindependently, whereas the remainder were recruited by the investigators after referrals to the dermatology (acne severity at 12 weeks), we analysed data according outpatient clinic. Four of 31 (13%) laser-treated patients to their original group assignment with an intention-to- withdrew, including two patients by 8 weeks and one by treat model and, for patients with missing data, used last 4 weeks, all three of whom left the locality. Another laser- available values as endpoint values. Additionally, we treated patient withdrew by 4 weeks after needing repeated this primary endpoint analysis twice, using first, systemic antibiotic treatment for worsening truncal acne.
a per protocol analysis that excluded all patients with One of ten controls withdrew by 4 weeks because of missing endpoint data and, second, an analysis that excluded only patients deemed to be missing completely Tables 1 and 2 show the baseline demographic and at random, whose group allocation was not thought to clinical characteristics of the two groups. Most patients have influenced their withdrawal from the trial. were young adults (38 of 41 were younger than 40 years, For acne severity, we did regression analysis to correct 31 of 41 were younger than 35 years) who had had a long for differences in baseline characteristics between the history of acne. Similar proportions of patients in each groups and to assess whether factors other than group group had previously received systemic antibiotics or oral allocation affected outcome. We fitted a forward stepwise isotretinoin. Most patients in both groups were white, multivariate regression model using the following and Asian and Afro-Caribbean ethnic groups were baseline characteristics that were judged capable of represented only in the laser treatment group. affecting outcome: age, sex, age of onset of acne, The difference between the groups’ acne severity at the duration of acne, skin type (Fitzpatrick classification 1–3 start of the trial was 0·2 on the Leeds revised grading vs 4–6), previous use of oral isotretinoin, and previous system. By intention-to-treat analysis, after 12 weeks, use of oral antibiotics for acne. On regression, any mean acne grade had improved from baseline by 1·9 (SD variable that was judged unimportant (p>0·25) was 1·8) in laser-treated patients and by 0·1 (SD 1·4) in discarded from the multivariate analysis. A repeatedmeasures analysis of variance was done to incorporate acne severity data from all timepoints and to explore value (r)
Use of systemic antibiotics for acne –0·2 Data for lesions are non-parametrically distributed, shown as median (IQR) and analysed per protocol. *Change from baseline after 12 weeks for laser vssham. †Data for severity are parametrically distributed, shown as mean (SD) Outcome is difference in Leeds revised acne grading score (12 weeks minus Table 2: Acne severity and lesion counts at baseline and Table 3: Regression analysis to assess effect of baseline characteristics on improvement of acne severity THE LANCET • Vol 362 • October 25, 2003 • www.thelancet.com *†Symptoms were seen in same patients. individual patients at the start and that at the end of the trial. The figure indicates that improvements were seen inlaser-treated patients who had a wide range of initial Figure 2: Change in acne severity with time severities and included those with severe disease, such as two patients with initial severity score of 7 and final scoreof 1, and one with initial severity score of 6 and final sham-treated patients (p=0·007) (table 2). Per protocol score of 1. After 12 weeks, severity had improved analysis and an analysis that excluded three laser-treated (reduced by at least 1 point on the grading scale) in 25 of patients who moved from the locality and were assumed 27 laser-treated patients and two of nine controls to be missing completely at random, produced similar (p=0·0001). Severity improved by at least 2 points in 16 results to the intention-to-treat analysis (reduction of of 27 patients treated with laser and in none of 9 controls acne grade from baseline for per protocol analysis, laser mean 2·1 [SD 1·5], sham 0·1 [SD 1·2], p=0·001; for Adverse events are shown in table 4. Six laser-treated analysis excluding patients missing completely at patients and two controls reported side-effects during the random, laser 2·1 [SD 1·6], sham 0·1 [SD 1·4], trial period. Two of the 31 patients who received laser treatment had deeply pigmented Afro-Caribbean skin.
Forward stepwise regression analysis failed to identify Both had moderate transient discomfort during any factor other than treatment-group allocation that irradiation at a high fluence (3·0 J/cm2), and one substantially affected the reduction in acne severity by described purpura that lasted 6 days on the side of the 12 weeks (table 3). Figure 2 shows the observed mean face that had been treated at this fluence. Three in the change in overall acne severity throughout the trial. Acne laser group and two controls reported short-term severity improved at every assessment in the laser- treatment group, the most rapid improvement occurring Total, inflammatory, and non-inflammatory lesion in the first 4 weeks after treatment. Repeated measures counts were similar in both groups at the start of the trial analysis of acne severity that used all data obtained at (table 2). After 12 weeks, a greater improvement in total every timepoint of the trial indicated a change in acne and inflammatory lesion counts was recorded in the laser- severity with time (p=0·001) and also with interaction treated group (table 2 and figure 3) than in controls.
between time and group allocation (p=0·001), and a Total lesion counts fell by 53% in laser-treated patients slight difference between treatment groups (p=0·09).
and by 9% in placebo-treated patients (p=0·023).
Figure 3 shows the relation between acne severity for Inflammatory lesion counts fell by 49% in laser-treated Acne severity at baseline (revised Leeds grade) Figure 3: Acne severity (A) and total lesion count (B) at baseline and 12 weeks*One control with multiple non-inflammatory lesions and a consequent high lesion count was excluded.
THE LANCET • Vol 362 • October 25, 2003 • www.thelancet.com a major concern.23 The proportion of acne patients carrying strains of P acnes resistant to tetracycline,erythromycin, or clindamycin rose from 34·5% to 64% between 1990 and 1997 in an urban population in the UK.24 At a time when prudent antibiotic prescription isbeing advocated in public-health initiatives to prevent the development of widespread global antibiotic resistance, the routine use of long courses of antibiotic treatment foracne should be re-assessed.25,26 Oral isotretinoin, a synthetic retinoid with powerful effects on cellular differentiation and division, is the most Data are non–parametrically distributed, expressed as median (IQR) and effective treatment and induces long-term remissions in a analysed per protocol. *To test whether change in acne severity after 12 weeks proportion of patients. Indications for its use have (from baseline) is similar for laser treatment with fluences of 3·0 and 1·5 J cm2.
recently broadened from nodulocystic acne to less severe Table 5: Effect of laser fluence on half-face lesion counts at forms, including mild-to-moderate disease that does not respond to systemic antimicrobials and acne associatedwith severe psychological problems.27 and by 10% in sham-treated patients (p=0·024). A trend isotretinoin causes dryness of the skin and mucous towards a reduction of non-inflammatory lesions membranes in most patients and has been associated with (comedones) was noted in patients treated by laser. Total more serious adverse events including: myalgia, lesion counts fell by at least 50% in 13 of 27 laser-treated arthralgia, benign intracranial hypertension, hepatitis, patients and in none of nine sham-treated patients hyperlipidaemia, acne fulminans, and visual 16 of the 31 laser-treated patients were randomised to isotretinoin and depression, suicide, psychosis, and receive treatment at fluences of 1·5 J/cm2 on the right side violent behaviour has recently been added to product of the face and 3·0 J/cm2 on the other side; the remaining information and remains under investigation.28 In the patients received 3·0 J/cm2 on the left and 1·5 J/cm2 on USA, despite awareness of the high teratogenicity of the right. Paired analysis failed to identify a significant isotretinoin and the implementation of strict guidelines difference between the change in lesion counts at these governing its prescription to women, about three pregnancy exposures take place per 1000 prescriptions ofthe drug.28,29 An optimum acne treatment would have longlasting Our results showed that PDL treatment was very well effectiveness in the control of active disease, improve tolerated in individuals with inflammatory acne. Afro- acne scarring, have few local or systemic side-effects, and Caribbean patients in our trial reported moderate transient discomfort during treatment with a fluence of The patients in our trial are likely to have been broadly 3·0 J/cm2, and one of these patients probably developed representative of adults with acne in the general transient purpura. These patients tolerated treatment at population, although recruitment of volunteers might 1·5 J/cm2 and responded well, suggesting that low laser have introduced a selection bias towards those with fluences should be used for deeply pigmented skin.
longstanding acne that had failed conventional However, this observation is based on two patients only, treatments. Masked studies are difficult to undertake with ablative lasers because the immediate development 12 weeks after one session of PDL treatment, acne of visible skin changes or pain can severely hinder severity improved substantially. This improvement was masking. However, low fluence non-ablative PDL seen for a range of disease severity, and included striking treatment usually produces no immediately obvious responses in three patients with severe acne. The changes to the skin. Two patients who had discomfort reduction in severity was indicated by a corresponding fall during treatment and communicated their experience to in total and inflammatory lesion counts, with almost 50% investigators might have introduced bias by suggesting of laser-treated patients and no controls having their total their treatment allocation to investigators. Since the lesion count halved by 12 weeks. The rapidity of the remaining patients reported no symptoms during response to laser treatment contrasts with that to treatment with non-ablative PDL, treatment allocation conventional treatments such as oral antibiotics, that often need administration for 6–8 weeks before benefits We used an intention-to-treat analysis for assessment are seen. The duration of the response suggested that the of acne severity and to carry forward last available results laser affected not only P acnes, as might be the mode of in missing patients, thereby maintaining the benefits of action of other light sources, which raises the possibility randomisation. Repeated analyses per protocol that that laser light might alter acne patients’ excluded these patients yielded similar results, suggesting immunobiological response to the bacterium. The that this approach was reasonable. Every laser-treated optimum treatment should be established by investigation patient in the study received treatment at two different of the effect of multiple treatments with a long follow-up fluences (1·5 J/cm2 and 3·0 J/cm2) on opposite sides of the and the mechanism of the therapeutic effect. midline to allow subgroup analysis of the effect of dosage.
Present acne treatments have several shortcomings.
Therefore, our primary outcome analysis actually Topical preparations are often irritating, cosmetically compares patients treated at two different fluences with unacceptable, and can bleach clothing or hair if they those who received no laser treatment. However, we believe this comparison to be appropriate in view of the effective, but responses to treatment are typically slow, absence of clinical difference between these fluences, and and continuous treatment for 6 to 8 months is usually because comparison of change in half-face lesion counts needed.21 Antibiotic-resistant strains of P acnes in patients between laser-treated and sham-treated patients (data treated for acne were first identified in 1979, and are now not shown) identified no substantial difference.
THE LANCET • Vol 362 • October 25, 2003 • www.thelancet.com PDLs are also reported to be effective in the treatment 10 Melo TB. Uptake of protoporphyrin and violet light photodestruction of atrophic acne scarring and reduced mean scar depth by of Propionibacterium acnes. Z Naturforsch [C] 1987; 42: 123–28.
48% after just one treatment.16 Our results suggest that 11 Papageorgiou P, Katsambas A, Chu AC. Phototherapy with blue (415nm) and red (660nm) light in the treatment of acne vulgaris. this laser treatment could be developed as a new Br J Dermatol 2000; 142: 973–78.
therapeutic approach that would allow simultaneous 12 Stillman S, Geen S, Harth Y, Shalita AR. High intensity narrow band treatment of both active acne and associated scarring. We blue light is effective in the treatment of acne vulgaris: an in vitro and believe that laser treatment should be further explored as in vivo study. J Eur Acad Dermatol Venereol 2000; 14 (suppl 1): 47.
an adjuvant or alternative to daily conventional 13 Tan OT, Sherwood K, Gilchrest BA. Treatment of children with port wine stains using the flashlamp-pumped tunable pulsed dye laser. N Engl J Med 1989; 320: 416–21.
14 Bjerring P, Clement M, Heickendorff L, Lybecker H, Kiernan M.
Dermal collagen production following irradiation by dye laser and R M Clement conceived the study. E D Seaton and A C Chu did most of broad-band light source. J Cosmet Laser Ther 2002; 3: 39–43.
the design study, data collection, treatment intervention, and result 15 Bjerring P, Clement M, Heickendorff L, Egekvist H, Kiernan M.
assessment. E D Seaton wrote the report and A C Chu revised the Selective non-ablative wrinkle reduction by laser. J Cut Las Ther manuscript. I Grace provided statistical advice and analysis. A Charakida 2000; 2: 9–15.
assisted with data collection, administration, and treatment intervention.
16 Patel N, Clement M. Selective nonablative treatment of acne scarring P E Mouser assisted with data collection and administration. with 585nm flashpump pulsed dye laser. Dermatol Surg 2002; 28:
17 O’Brien SC, Lewis JB, Cunliffe WJ. The Leeds revised acne grading Since completion of the trial, EDS has started laboratory research into system. J Dermatol Treatment 1998; 9: 215–20.
the mechanism of action of PDL therapy in acne at the Department of 18 Lehmann HP, Robinson KA, Andrews JS, Holloway V Goodman Dermatology, Imperial College, London, UK, with financial support SN. Acne therapy: a methodologic review. J Am Acad Dermatol 2002; from EUPhotonics. RMC was an academic employee of EUPhotonics, 47: 231–40.
and contributed to development of the laser and trial conception, but not 19 Lucky AW, Barber BL, Girman CJ, Williams J, Ratterman J, to detailed trial design, data collection, data analysis, or interpretation of Waldstreicher J. A multirater validation study to assess the the results. The other authors have no conflict of interest.
reliability of acne lesion counting. J Am Acad Dermatol 1996; 35:
20 Altman DG. Practical statistics for medical research. London: EUPhotonics (Swansea, Wales) provided the laser. We thank Michael Kiernan, Jay Birnbaum, and Douglas Seaton for very helpful 21 Greenwood R, Burke B, Cunliffe WJ. Evaluation of a therapeutic comments during the study, and Jean Berkeley and Jen Morris for strategy for the treatment of acne vulgaris with conventional therapy.
Br J Dermatol 1986; 114: 353–58.
22 Olsen TE. Therapy of acne. Med Clin North Am 1982; 66:
23 Crawford WW, Crawford IP, Staughton RB, Cornell RC. Laboratory Jowett S, Ryan T. Skin disease and handicap: an analysis of the induction and clinical occurrence of combined clindamycin and impact of skin conditions. Soc Sci Med 1985; 20: 425–29.
erythromycin resistance in Corynebacterium acnes. J Invest Dermatol Gupta MA, Gupta AK. Depression and suicidal ideation in 1979; 72: 187–90.
dermatology patients with acne, alopecia areata, atopic dermatitis and 24 Coates P, Vyakrnam S, Eady EA, Jones CE, Cove JH, Cunliffe WJ.
psoriasis. Br J Dermatol 1998; 139: 846–50.
Prevalence of antibiotic-resistant propionibacteria on the skin of acne Cotterill JA, Cunliffe WJ. Suicide in dermatological patients. patients: 10-year surveillance data and snapshot distribution study. Br J Dermatol 1997; 137: 246–50.
Br J Dermatol 2002; 146: 840–48.
Cunliffe WJ. Unemployment and acne. Br J Dermatol 1986; 115: 386.
25 Center for Disease Control and Prevention, Food and Drug Norris JF, Cunliffe WJ. A histological and immunocytochemical Administration, and National Institutes of Health. A public health study of early acne lesions. Br J Dermatol 1988; 118: 651–59.
action plan to combat antimicrobial resistance: part 1—domestic Mouser PE, Baker BS, Seaton ED, Chu AC. Propionibacterium issues. http://www.cdc.gov/drugresistance/actionplan/aractionplan.pdf acnes reactive Th-1 cells in the skin of patients with acne vulgaris. J Invest Dermatol (in press).
26 Department of Health. Antimicrobial resistance strategy and action Leyden JL, McGinley KJ, Mills OH, Kligman AM.
plan. London: Department of Health, 2000: 23.
Propionibacterium levels in patients with and without acne vulgaris. 27 Ortonne JP. Oral isotretinoin treatment policy: do we all agree? J Invest Dermatol 1975; 65: 382–84.
Dermatol 1997; 195 (suppl 1): 34–37.
Eady EA, Jones CE, Tipper JL, Cove JH, Cunliffe WJ, Layton AM.
28 US Food and Drug Administration, Center for Drug Evaluation and Antibiotic resistant propionibacteria in acne: need for policies to Research. Accutane label (revised) 2002 drug information.
modify antibiotic usage. BMJ 1993; 306: 555–56.
http://www.fda.gov/cder/foi/label/2002/18662s051lbl.pdf (accessed Kjeldstad B, Johnsson A. An action spectrum for blue and near ultraviolet inactivation of Propionibacterium acnes; with emphasis on 29 Mitchell AA, van Bennekom CM, Louik C. A pregnancy prevention a possible porphyrin photosenstisation. Photochem Photobiol 1986; 43:
program in women of childbearing age receiving isotretinoin. N Engl J Med 1995; 333: 101–06.
THE LANCET • Vol 362 • October 25, 2003 • www.thelancet.com
Effects of childhood body size on breast cancer tumour characteristics Jingmei Li1,2, Keith Humphreys1, Louise Eriksson1, Kamila Czene1, Jianjun Liu2 and Per 1Karolinska Institutet, Department of Medical Epidemiology and Biostatistics, Box 281, 171 2Human Genetics, Genome Institute of Singapore, 60 Biopolis Street, Singapore, 138672, Corresponding author: Jingmei Li, Jingmei.Li@ki.se A