Volume 174 Issue 10

Indigenous Health Research Childhood post-streptococcal glomerulonephritis as a risk factor for chronic renal disease in later life Andrew V White, Wendy E Hoy and David A McCredie MJA 2001; 174: 492-496 For editorial comment, see Atkins Abstract - Methods - Results - Discussion - Acknowledgements - References - Authors' details - - More articles on Aboriginal health Abstract Objective: To test the hypothesis that post-streptococcal glomerulonephritis (PSGN) in childhood is a risk factor for chronic renal disease in later life. Design: Retrospective cohort study. Setting: A remote Aboriginal community in the "Top End" of the Northern Territory that experienced two epidemics of PSGN in 1980 and 1987, respectively. Participants: 472 people who were aged 2-15 years during either epidemic. They were categorised by clinical features recorded during the epidemics as having clinically defined PSGN (63), "abnormal urine" (haematuria or proteinuria; 86) or controls (323). Outcome measures: Urinary albumin to creatinine ratio (ACR), haematuria (by dipstick urinalysis), blood pressure, serum creatinine level, and calculated glomerular filtration rate (GFR) during community screening in 1992-1998. Results: Overt albuminuria (ACR > 34 mg/mmol) was present at follow-up in 13% of the PSGN group, 8% of the abnormal urine group, and 4% of the control group. The odds ratio (OR) for overt albuminuria in those with a history of PSGN compared with the control group, adjusted for age and sex, was 6.1 (95% CI, 2.2-16.9). Haematuria (> trace) was present in 21% of the PSGN group compared with 7% of the control group (adjusted OR, 3.7; 95% CI, 1.8-8.0). There were no significant differences between the groups in blood pressure, serum creatinine level or calculated GFR. Conclusion: In this population, a history of PSGN in childhood is a risk factor for albuminuria and haematuria in later life. Although unusual in the rest of Australia, post-streptococcal glomerulonephritis (PSGN) is still common in Aboriginal children living in remote communities, where group A streptococcal pyoderma is endemic.1 In these communities, chronic renal disease and end-stage renal failure also occur in adults at alarming rates.2-4PSGN is usually followed by clinical recovery over several days to weeks, and the long-term outlook has generally been regarded as excellent, with no increase in risk of urinary abnormalities or hypertension.5-7 However, some studies have suggested an increase in rates of chronic renal impairment after this illness.8,9 We aimed to test whether a history of PSGN in childhood is a risk factor for later renal dysfunction in Aboriginal Australians living in a remote community. The main outcome measure used, albumin to creatinine ratio, is a sensitive early marker of renal damage. It has been shown to provide a reliable estimate of 24-hour protein excretion and to predict the rate of decline of glomerular filtration rate and progression to end-stage renal failure in diabetic10 and non-diabetic11 nephropathy. Albuminuria has also been shown to mark early chronic renal disease in this population of Aboriginal Australians, and its progression predicts renal failure, as well as cardiovascular disease and mortality.12,13 Methods Study design This was a retrospective cohort study of children from an isolated Aboriginal coastal community in the "Top End" of the Northern Territory of Australia. The community experienced two epidemics of PSGN in 1980 and 1987, respectively, each lasting for three months.14,15 Children were followed up after these epidemics for a mean of 14.6 years (range, 6-18 years). The study was approved by the Joint Institutional Ethics Committee of the Royal Darwin Hospital and the Menzies School of Health Research, as well as a local community health board. Consent was obtained from each individual or guardian at the time of screening. Participants Participants were 472 people who lived in the community and were aged 2-15 years at the time of either epidemic and who participated in health screening examinations between 1992 and 1998. These 472 people represented 98% of the population of the community in the relevant age groups, according to 1996 census estimates.16 Baseline data During the epidemics, children in the community were screened systematically for oedema, hypertension, and urinary abnormalities on dipstick testing; results were recorded in individuals' medical records in the community. We used these data to categorise children by history during the epidemics: The PSGN group had documented oedema (facial swelling or dependent oedema) or hypertension (diastolic pressure ≥ 80 mmHg if aged 2-12 years and ≥ 85 mmHg if over 12, levels corresponding to the 90th percentile for each age range)17 plus haematuria greater than trace or proteinuria greater than trace on dipstick urinalysis. The "abnormal urine" group had haematuria greater than trace or proteinuria greater than trace, but no oedema or hypertension. The control group comprised children who had normal results on clinical examination and urinalysis (trace or less for blood and protein); children with no symptoms suggesting PSGN, but for whom urinalysis was either not performed or not recorded; and children with no entry in the medical record at the time of the epidemics. Children whose ages were in the range 2-15 years during both epidemics were categorised according to their most abnormal findings in either epidemic. Outcome measures Population health screening was undertaken in the community between 1992 and 1999.14 For people screened more than once, results from the latest screening were used. The albumin to creatinine ratio (ACR) was determined in a random urine sample and was categorised as normal (< 1.1 mg/mmol), suspicious (1.1-3.3 mg/mmol), microalbuminuria (3.4-33 mg/mmol), or overt albuminuria (≥ 34 mg/mmol). Glomerular filtration rate (GFR) was calculated using the formula of Cockroft and Gault.18 Dipstick urinalysis was also performed (Multistix 10SG, Bayer Diagnostics), and blood pressure and serum creatinine level were measured. Analysis Baseline characteristics of the groups were compared using Fisher's exact test for categorical variables and, as not all data followed a normal distribution, the non-parametric Kruskal-Wallis test for continuous variables. Logistic regression estimates were used to obtain adjusted proportions of the population with albuminuria. Odds ratios for the outcomes albuminuria and haematuria were obtained from logistic regression models that included the factors age, sex, birth weight and body mass index. Analyses were performed using Stata statistical software.19Results Baseline characteristics Of the 472 people included in the study, 259 were aged 2-15 years during the 1980 epidemic, and 331 during the 1987 epidemic (with 118 in the age group during both epidemics). Overall, 275 (58%) were male. Categorisation of participants Categorisation of participants according to history during the epidemics is shown in Box 1. Of the 63 children with clinically defined PSGN, all had haematuria and proteinuria, 61 (97%) had oedema and 28 (44%) had hypertension. Although evidence of preceding group A streptococcal infection was not required for classification in the PSGN group, serum antideoxyribonuclease B antibody titres were positive (≥ 1:480) in all 36 of the group in whom they were measured, and serum complement levels were consistent with PSGN (low C3 level) in 35 of the 39 in whom they were measured. Of the 86 participants in the abnormal urine group, 84 (98%) had haematuria, and 24 (28%) had proteinuria. Characteristics of participants at follow-up differed significantly between the groups, with the PSGN group being younger, and the abnormal urine group having a lower proportion of males (Box 2). The three control subgroups also differed at follow-up in median age (normal results subgroup, 26.9 years; no urinalysis subgroup, 17.5 years; and not recorded subgroup, 18.4 years; P = 0.001) and body mass index (normal results subgroup, 20.9 kg/m2; no urinalysis subgroup, 19.1 kg/m2; and not recorded subgroup, 19.8 kg/m2; P = 0.007). However, after adjustment for age and sex, there were no significant differences in height or weight. The control subgroups were combined for analysis. Outcomes On follow-up screening, 104 participants (22%) had albuminuria of any degree (micro- or overt; ACR ≥ 3.4 mg/mmol), 27 (6%) had overt albuminuria (ACR ≥ 34 mg/mmol), and 45 (10%) had haematuria (≥ trace), while 64 (14%) had haematuria or overt albuminuria. Albuminuria and haematuria were more prevalent in the groups with a history of clinical PSGN or abnormal urine during the PSGN epidemics than in the control group (Box 3). In the abnormal urine group, outcomes at follow-up were similar whether or not haematuria had occurred alone or in the presence of proteinuria during the epidemics. As the presence of albuminuria is significantly related to age in this community,14 and as albuminuria was more common in females than males (overt albuminuria occurred in 4% of males and 8% of females; P = 0.044), probabilities were adjusted for age and sex (Box 4). The adjusted probability of overt albuminuria at follow-up was 13.6% after PSGN (95% CI, 6.7%-25%), compared with 2.5% in controls (95% CI, 1.1%-4.8%). Odds ratios for albuminuria and haematuria at follow-up according to history during the epidemics are shown in Box 3. After adjustment for age and sex, the odds of overt albuminuria were more than six times greater after PSGN compared with the control group, while the odds of albuminuria of any degree were more than three times greater. The population-attributable fraction, or proportion of overt albuminuria in the study population that can be attributed to PSGN in childhood, was 24% (95% CI, 5%-40%). After adjustment for age and sex, the odds of haematuria were more than three times greater after PSGN compared with the control group, while the odds of either haematuria or overt albuminuria were five times greater. Only three individuals had both haematuria and overt albuminuria, two of whom had a history of PSGN. Birth weight was available for 429 participants (61 with PSGN, 77 with abnormal urine and 291 controls). Adding birth weight to the logistic regression model gave an odds ratio of 7.6 for overt albuminuria in the PSGN group using controls as the reference (95% CI, 2.5-22.5). Adding body mass index to the model did not significantly alter the odds ratios. There were no significant differences in blood pressure, serum creatinine level or calculated glomerular filtration rate between the groups. Discussion This study indicates that a history of PSGN in childhood is a risk factor for albuminuria and haematuria years later, and suggests that about a quarter of cases of overt albuminuria may be attributable to PSGN in childhood. The incidence of renal disease is high in this population of Aboriginal Australians,12 and other risk factors for renal disease are also common, including low birth weight, recurrent infectious diseases, diabetes and features of syndrome X.3,12,20,21 Possibly, it is the combination of insults that leads to high risk for later renal disease. A prospective cohort study would provide the best evidence. Results of other studies on the contribution of PSGN to chronic renal disease have varied, with some studies reporting no link. For example, two large follow-up studies after epidemic PSGN in Trinidad6,22,23 and Venezuela,7,24 respectively, reported low rates of long-term abnormalities, although the Venezuelan study found that 11.2% of participants had proteinuria of > 500 mg/24 h at 11-year follow-up. These studies had high losses to follow-up (31% and 82%, respectively), and neither had a control group nor assessed microalbuminuria. A cohort study after a PSGN epidemic in an American Indian community found no difference at 10-year follow-up between those who had had PSGN and those who had not in blood pressure, serum creatinine level, proteinuria or haematuria. However, urinary abnormalities were common, being present in 17% of the PSGN group and 13% of the control group.5 In contrast, other studies have reported, similarly to ours, clinically important abnormalities at long-term follow-up after PSGN. An uncontrolled study from north India found proteinuria (defined as more than trace levels on qualitative examination) in 13.8% of people two to 10 years after nephritis.25 Protein to creatinine ratios were > 20 mg/mmol in 9% of people recruited from a tertiary London hospital 14-22 years after sporadic childhood PSGN,26 while, in an Italian study, microalbuminuria or greater was present in 46% of 26 patients three to 24 years after PSGN but only 2.5% of 100 control participants.27 Other studies have found significantly lower renal functional reserve in people with a remote history of PSGN compared with control participants.28,29 Our study had the strengths of having a control group and a long follow-up, studying a large proportion of people in a single community and measuring albuminuria in the microalbuminuric range. Its limitations include possible misclassification of participants, as PSGN was diagnosed by clinical criteria. The abnormal urine group may have included people with subclinical PSGN, other renal disease, or isolated haematuria of no significance. However, this is unlikely to have biased results significantly, as the main findings concerned differences between the PSGN and control groups, which had more certain definitions. Nevertheless, some participants classified with PSGN may have had another cause for their renal disease. For example, one child was later diagnosed with mesangiocapillary glomerulonephritis after renal biopsy; she may have been predisposed to PSGN by pre-existing renal disease, may have had consecutive disease processes or may never have had PSGN. Lastly, control participants who were not seen during the epidemics may have had unrecognised PSGN. However, this would have decreased rather than exaggerated differences between the PSGN and control groups. Although some historical data were unavailable, fewer females than males were studied, and follow-up times varied; these factors were unlikely to have affected results. Because albuminuria precedes clinical signs of chronic renal disease, longer follow-up could be expected to show changes in blood pressure, serum creatinine levels and GFR. Other factors may be involved in the observed relationship between PSGN and albuminuria and the postulated relationship with chronic renal disease. Firstly, another underlying renal process may predispose both to PSGN on exposure to a nephritogenic streptococcus and to later albuminuria. Secondly, albuminuria may not have such adverse prognostic significance after PSGN as it does in other circumstances; our follow-up was not long enough to show progression to chronic renal disease. Thirdly, PSGN may increase risk of chronic renal disease only in combination with other insults. Renal disease is extremely common in this community, and, although our findings are likely to apply to similar populations, they may not be universally applicable. In summary, we have presented evidence that, in this community, a remote history of PSGN in childhood is a powerful risk factor for renal damage, as evidenced by increased ACR and haematuria. These findings are important as PSGN is still prevalent in children living in Aboriginal communities in Australia. Prevention of PSGN is possible through improvements to housing, economic and living conditions, along with attention to control and treatment of scabies and skin infections. Preventing PSGN may contribute to reducing the incidence of renal disease and renal failure in the future. Acknowledgements This study was supported by the National Health and Medical Research Council and the Australian Kidney Foundation. We acknowledge the support and participation of the Tiwi community and the staff of the health clinic at Nguiu. Health workers Jerome Kerinauia, Nellie Punguatji, Darren Fernando and Colleen Kantilla and project officers Eric and Elizabeth Tipiloura were key contributors to the field work. We thank Bev Hayhurst, who coordinated much of the screening program, and Zhiqiang Wang, who provided statistical advice. We also acknowledge Kate Walker's work in looking at earlier data. References Streeton CL, Hanna JN, Messer RD, Merianos A. An epidemic of acute post-streptococcal glomerulonephritis among Aboriginal children. J Paediatr Child Health 1995; 31: 245-248. Spencer JL, Silva DT, Snelling P, Hoy WE. An epidemic of renal failure among Australian Aboriginals. Med J Aust 1998; 168: 537-541. Hoy WE, Norman RJ, Hayhurst BG, Pugsley DJ. A health profile of adults in a Northern Territory aboriginal community, with an emphasis on preventable morbidities [see comments]. Aust N Z J Public Health 1997; 21: 121-126. Cass A, Gillin AG, Horvath JS. End-stage renal disease in Aboriginals in New South Wales: a very different picture to the Northern Territory. Med J Aust 1999; 171: 407-410. Perlman L, Herdman R, Kleinman H, Vernier R. Poststreptococcal glomerulonephritis. A ten year follow up of an epidemic. JAMA 1965; 194: 63-70. Potter E, Lipschultz S, Abidh S, et al. Twelve- to seventeen-year follow up of patients with poststreptococcal acute glomerulonephritis in Trinidad. N Engl J Med 1982; 307: 725-729. Garcia R, Rubio L, Rodriguez-Iturbe B. Long-term prognosis of epidemic poststreptococcal glomerulonephritis in Maracaibo: follow-up studies 11-12 years after the acute episode. Clin Nephrol 1981; 15: 291-298. Baldwin DS, Gluck MC, Schacht RG, Gallo G. The long-term course of poststreptococcal glomerulonephritis. Ann Intern Med 1974; 80: 342-358. Schacht RG, Gallo GR, Gluck MC, et al. Irreversible disease following acute poststreptococcal glomerulonephritis in children. J Chronic Dis 1979; 32: 515-524. Rodby RA, Rohde RD, Sharon Z, et al. The urine protein to creatinine ratio as a predictor of 24-hour urine protein excretion in type 1 diabetic patients with nephropathy. The Collaborative Study Group. Am J Kidney Dis 1995; 26: 904-909. Ruggenenti P, Gaspari F, Perna A, Remuzzi G. Cross sectional longitudinal study of spot morning urine protein:creatinine ratio, 24 hour urine protein excretion rate, glomerular filtration rate, and end stage renal failure in chronic renal disease in patients without diabetes [published erratum appears in BMJ 1998; 317: 1491]. BMJ 1998; 316: 504-509. Hoy WE, Mathews JD, McCredie DA, et al. The multidimensional nature of renal disease: rates and associations of albuminuria in an Australian Aboriginal community. Kidney Int 1998; 54: 1296-1304. McDonald S, Wang Z, Hoy WE. Physical and biochemical predictors of death in an Australian Aboriginal cohort. Clin Exp Pharmacol Physiol 1999; 26: 618-621. Gogna NK, Nossar V, Walker AC. Epidemic of acute poststreptococcal glomerulonephritis in aboriginal communities. Med J Aust 1983; 1: 64-66. Devanesen D, Bernard E, Stokes M, et al. Lessons from an outbreak of glomerulonephritis in an aboriginal community. Annual report 1987-88, Menzies School of Health Research. Darwin: Menzies School of Health Research, 1988. Australian Bureau of Statistics. 1996 census of population and housing: Basic community profile. Canberra: ABS, 1996. (Catalogue no. 2020.0.) Report of the Second Task Force on Blood Pressure Control in Children 1987. Task Force on Blood Pressure Control in Children. National Heart, Lung, and Blood Institute, Bethesda, Maryland. Pediatrics 1987; 79: 1-25. Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976; 16: 31-41. Stata Statistical Software [program]. Release 6.0. College Station, Texas: Stata Corporation, 2000. Hoy WE. Renal disease in Australian aboriginals [editorial]. Med J Aust 1996; 165: 126-127. Hoy WE, Rees M, Kile E, et al. A new dimension to the Barker hypothesis: low birthweight and susceptibility to renal disease. Kidney Int 1999; 56: 1072-1077. Potter E, Abidh S, Sharrett R, et al. Clinical healing two to six years after poststreptococcal glomerulonephritis in Trinidad. N Engl J Med 1978; 298: 767-772. Nissenson A, Mayon-White R, Potter E, et al. Continued absence of clinical renal disease seven to 12 years after poststreptococcal acute nephritis in Trinidad. Am J Med 1979; 67: 255-262. Rodriguez-Iturbe B, Garcia R, Rubio L. Epidemic glomerulonephritis in Maracaibo. Evidence for progression to chronicity. Clin Nephrol 1976; 5: 197-205. Singhal PC, Malik GH, Narayan G, et al. Prognosis of post-streptococcal glomerulonephritis: Chandigarh study. Ann Acad Med Singapore 1982; 11: 36-41. Clark G, White RH, Glasgow EF, et al. Poststreptococcal glomerulonephritis in children: clinicopathological correlations and long-term prognosis. Pediatr Nephrol 1988; 2: 381-388. Buzio C, Allegri L, Mutti A, et al. Significance of albuminuria in the follow-up of acute poststreptococcal glomerulonephritis. Clin Nephrol 1994; 41: 259-264. Rodriguez-Iturbe B, Herrera J, Garcia R. Response to acute protein load in kidney donors and in apparently normal postacute glomerulonephritis patients: evidence for glomerular hyperfiltration. Lancet 1985; 2: 461-464. Cleper R, Davidovitz M, Halevi R, Eisenstein B. Renal functional reserve after acute poststreptococcal glomerulonephritis. Pediatr Nephrol 1997; 11: 473-476. (Received 5 Jul 2000, accepted 7 Dec 2000) Authors' details Menzies School of Health Research, Darwin, NT. Andrew V White, FRACP, Research Student, Menzies School, and Flinders University NT Clinical School, Darwin NT; currently, Paediatrician, Remote Health, Alice Springs, NT. Wendy E Hoy, FRACP, Principal Research Fellow. Royal Children's Hospital, Melbourne, VIC. David A McCredie, MD, FRACP, Nephrologist. Reprints will not be available from the authors. Correspondence: Dr A V White, Remote Health Services, PO Box 721, Alice Springs, NT 0871. Andrew. WhiteATnt.gov.au Make a comment Back to text 2: Characteristics of participants at follow-up, by diagnostic category during post-streptococcal glomerulonephritis (PSGN) epidemics PSGN (n = 63) Abnormal Control (n = 86) Urine (n = 323) P % Male 65% 45% 60% 0.02* Median age in years (range) 18.1 (8-31) 23.4 (13-32) 19.0 (10-33) 0.001† Median BMI (kg/m2) (range) 18.8 (13-45) 20.0 (15-47) 19.8 (13-40) 0.221† BMI = body mass index. * Fisher's exact test. †Kruskal-Wallis test. Back to text 3: Outcomes on follow-up screening, by diagnostic category during post-streptococcal glomerulonephritis (PSGN) epidemics Outcome PSGN (n = 63) Abnormal urine (n = 86) Control (n = 323) ACR ≥ 34 mg/mmol Rate 13% 8% 4% Crude odds ratio (95% CI) 3.8 (1.5-9.8) 2.3 (0.88-6.0) 1 Adjusted* odds ratio (95% CI) 6.1 (2.2-16.9) 1.6 (0.6-4.2) 1 PAF (95% CI) 24% (5%-40%) 8% (-14% to 26%) ACR ≥ 3.4 mg/mmol Rate 32% 30% 18% Crude odds ratio (95% CI) 2.2 (1.2-4.0) 2.0 (1.2-3.4) 1 Adjusted* odds ratio (95% CI) 3.2 (1.7-6.2) 1.4 (0.8-2.6) 1 PAF (95% CI) 11% (4%-18%) 5% (-4% to 14%) Haematuria > trace Rate 21% 9% 7% Crude odds ratio (95% CI) 3.4 (1.6-7.2) 1.4 (0.6-3.1) 1 Adjusted* odds ratio (95% CI) 3.7 (1.8-8.0) 1.1 (0.4-2.6) 1 PAF (95% CI) 20% (5%-33%) 1% (-14% to 14%) Haematuria > trace or ACR ≥ 34 mg/mmol Rate 30% 16% 11% Crude odds ratio (95% CI) 3.6 (1.9-6.8) 1.6 (0.8-3.1) 1 Adjusted* odds ratio (95% CI) 4.6 (2.3-9.0) 1.2 (0.6-2.4) 1 PAF (95% CI) 19% (8%-29%) 3% (-9% to 13%) ACR = albumin to creatinine ratio. PAF = population-attributable fraction. * Adjusted for age and sex. Back to text Back to text

Andrew V White · Wendy E Hoy · David A McCredie

Public Health Has the ban on smoking in New South Wales restaurants worked? A comparison of restaurants in Sydney and Melbourne Simon Chapman, Ron Borland and Anita Lal MJA 2001; 174: 512-515 Abstract - Methods - Results - Discussion - Acknowledgements - References - Authors' details - - More articles on Public and environmental health Abstract Objective: To evaluate compliance with a legislative ban on smoking inside restaurants by comparing smoking in Sydney restaurants (where it is legally banned) with smoking in Melbourne restaurants (not subject to a legal ban). Design and participants: Unobtrusive observational study of restaurant patrons, and interviews with restaurant staff, carried out by 159 volunteers. Setting: 78 Sydney restaurants with smoke-free indoor environments (as required by legislation) and 81 Melbourne restaurants not subject to legislation preventing smoking. The study took place from 20-31 October 2000. Intervention: Legislation to ban smoking in indoor areas of restaurants was introduced in New South Wales in September 2000 (about six weeks before our study). Outcomes: Observed incidents of smoking inside restaurants; staff attitudes to the ban; customer satisfaction as indicated by comments to staff; staff perceptions of restaurant patronage. Results: No restaurant patrons were seen smoking in 78 Sydney restaurants during 156 hours of observation of 2646 diners, compared with 176 smokers among 3014 Melbourne diners over 154 hours of observation. Thirty-one per cent (24/78) of Sydney restaurants had experienced smokers attempting to smoke indoors after the legislation was introduced; 6% (5/78) reported instances of smokers refusing to stop smoking when asked; 79% (62/78) of restaurants had received favourable comments from patrons about the smoke-free law; 81% (63/78) of restaurant staff interviewed either supported or strongly supported the law. Since introduction of the legislation, 76% of restaurants reported normal trade, 14% increased trade, and 9% reduced trade. Conclusions: Smoke-free restaurants do not require "smoking police" to enforce bans, present few ongoing difficulties for staff, attract many more favourable than unfavourable comments from patrons, and do not adversely affect trade. Throughout more than 15 years of advocacy for an Australia-wide ban on smoking in restaurants, opponents of smoke-free policies (principally the tobacco industry and the Australian Hotels Association1) have argued that such a ban would have serious negative consequences: patronage would be reduced; smoking tourists would choose other travel destinations; staff would be reluctant to enforce the ban; and without expensive government "smoking police" many smokers would ignore the ban.2,3Advocates of smoke-free restaurants have argued that bans not only bring public and occupational health benefit, but are very popular with diners4 and do not reduce trade.5-8 In June 2000, the New South Wales government introduced legislation to ban smoking inside all restaurants, cafes and other places where food is purchased and consumed indoors.9 The law took effect in September 2000, 10 days before the commencement of the Sydney Olympic Games. Smoking was still permitted in outdoor eating areas and in bar areas inside those restaurants that had a "reception" licence (ie, typically, a bar area where patrons are served drinks before a meal). The Victorian State government announced in May 2000 that a smoking ban, identical in most respects to that operating in New South Wales, would be introduced to take effect from 1 July 2001.10 Prior to the ban on indoor smoking in NSW restaurants, an increasing number of restaurants in both States had voluntarily introduced totally smoke-free dining, designated smoking areas or time restrictions (eg, smoking permitted only after 10 pm). The introduction of anti-smoking legislation in NSW prior to its introduction in Victoria afforded a unique window of opportunity to evaluate its implementation by comparing amount of smoking, staff acceptance of bans and customer numbers in restaurants in jurisdictions with and without a ban. Methods One hundred and fifty-nine volunteers (78 in Sydney, 81 in Melbourne), mainly employed by or studying at the University of Sydney or the VicHealth Centre for Tobacco Control, were recruited via email to be observers in our study. Over the period 20-31 October 2000, the observers were invited to dine in Sydney and Melbourne restaurants and were reimbursed $40 towards the cost of their meal. The restaurants (selected by us from dining guides) were restricted to locations within three kilometres of the suburb of residence of the volunteers. No attention was given in selection to the type of food served, the prices charged or whether the restaurant had banned smoking as restaurant policy before the legislative ban (in Sydney) or already, despite absence of a legal ban (in Melbourne). The restaurants selected in both Sydney and Melbourne were spread over a fairly broad geographical area. No restaurant was observed more than once. Observers, after familiarising themselves with a standard observer's protocol, were asked to go their allocated restaurant in the evening, request a table inside the restaurant, order a meal and eat at their leisure. During the evening they tallied the number of diners and patrons waiting to collect take-away meals in the same room as themselves; recorded any indoor smoking incidents and whether smoking occurred in any bar areas; noted whether ashtrays were present; and, before leaving, administered to their table waiter a brief eight-item questionnaire (designed by us) on experiences with the new law and attitudes toward smoking in restaurants. In Melbourne there were two different versions of the questionnaire, depending on whether a ban had already been implemented or whether staff were expressing expectations relating to a future ban. Any comments made by restaurant staff were also noted. Results Observation In Sydney, 78 restaurants were observed for a mean time of 120 minutes (range, 45-240 minutes). The mean number of diners per restaurant was 34 (range, 4-115). None of the 2646 diners in these restaurants were seen to smoke. Ashtrays were observed on tables in three restaurants (4%). In Melbourne, 81 restaurants were observed for a mean time of 114 minutes (range, 60-235 minutes). The mean number of diners per restaurant was 37 (range, 5-198). Of the 3014 patrons, 176 diners (6%) smoked in indoor eating areas during the observation period. Overall, 15% (12/81) of restaurants had voluntarily introduced total bans on smoking indoors, with another 7% (6/81) banning smoking in all eating areas, giving 22% (18/81) having management-initiated bans comparable to those in law in NSW. A further 30% (24/81) had bans in some areas or at certain times, leaving 48% (39/81) that did not provide some notionally smoke-free dining. Smoking practices were observed for compliance with individual restaurant policies on smoking (Box 1). It was noted that intrusion of smoke into smoke-free areas could be a problem, but there was no evidence of violation of management bans. Moreover, in areas where smoking was allowed, there were a considerable number of instances in which no smoking was actually observed. The presence of ashtrays in areas where smoking was allowed was associated with marginally more smoking -- smoking was observed in 20% (16/81) of restaurants with ashtrays versus 14% (11/81) of restaurants with no ashtrays. Staff interviews In Sydney, staff were asked about their experiences since the implementation of the mandated ban (Box 2). With the ban having been in place for about six weeks, 31% (24/78) of restaurants had experienced smokers attempting to smoke indoors since the ban, mostly in the first two weeks. Thirty-three per cent (26/78) had experienced at least one incident of smokers complaining or "making a scene" about not being allowed to smoke; however, only 6% (5/78) had experienced smokers refusing to stop smoking when asked. Most attempts to smoke occurred with people who were unaware of the new law. Sixteen restaurants (21%) reported that smokers, when told that they could not smoke indoors, had gone elsewhere before ordering. The frequency of this was not recorded, but no restaurant staff described it as a common occurrence. In contrast to the one-third of restaurants receiving complaints, 79% (62/78) of Sydney restaurants had received favourable comments from patrons about the ban, with over half receiving several or many comments since its implementation. Of restaurant staff who were questioned, 81% (63/78) supported the ban. Several of those who were against the ban commented that they were smokers themselves and resented being no longer able to smoke while working. While most businesses (76%) reported patronage in the current week to be normal, 14% of restaurants reported being busier than normal and 9% said business was slower. In Melbourne, in the 18 restaurants with eating area bans and seven of the restaurants with partial bans (25/81), staff were asked questions about implementation of bans. In restaurants with no bans, staff were asked about what they would expect to happen if bans were introduced (Box 2). In Melbourne restaurants, expectations and experiences of staff were generally similar, but staff in Melbourne observed or expected patrons to be more resistant to smoking bans than did staff in Sydney. It is notable that reports of patrons leaving without eating or saying they "won't be back" appear to be less frequent where bans are mandated. It is also notable that levels of positive comments were at least as high in Melbourne as Sydney. However, the small numbers preclude meaningful statistical analyses of differences. Discussion Our study shows that Sydney diners do not require "smoking police"11 to convince them not to smoke inside restaurants. Staff have dealt with smoking incidents by reminding patrons of the law, and only a small proportion of restaurants have experienced incidents of smokers refusing to put out their cigarettes when asked. While one in three restaurants reported incidents of smoking in the first weeks of the ban, compliance with the new law now appears to be exceptionally high. In Melbourne, while an encouraging number of restaurants have voluntarily introduced smoking bans, staff experience of people complaining, going elsewhere or generally "making a scene" appears to be more common than in Sydney, where the ban may have introduced a "level playing-field", whereby patrons threatening to "go elsewhere" find that all restaurants now have the same non-smoking conditions. Increasingly, Australian studies have demonstrated that a majority of restaurant patrons prefer to eat in situations where smoking is banned or meaningfully restricted,12 and that restaurateurs have for years severely underestimated the demand for smoke-free dining.13-16Economic game theory17 predicts that restaurant owners trying to introduce voluntary bans could be caught in a dilemma between complying with occupational health and safety laws and yielding to perceived competitive pressures to allow smoking. Under these circumstances, many would continue to put their (often erroneously) perceived economic interests first. Internal tobacco industry documents affirm that the industry has long been intensely concerned about the impact of smoking restrictions on tobacco sales.18 Just as smoke-free workplaces have dramatically reduced smoking frequency among continuing smokers,19 smoke-free dining also reduces smoking opportunities and so is likely to further reduce net tobacco consumption. The most immediate effect of such bans is to reduce the occupational exposure of restaurant staff to environmental tobacco smoke.20 Internal industry documents also reveal that the tobacco industry was aware that its public campaigns predicting dire economic consequences for restaurants were disingenuous. A 1994 Philip Morris internal document stated that "the economic arguments often used by the industry to scare off smoking ban activity were no longer working, if indeed they ever did. These arguments simply had no credibility with the public, which isn't surprising when you consider that our dire predictions in the past rarely came true."21 Our findings suggest that the Sydney restaurant trade has not declined, despite smoking bans — if anything, it may have increased slightly. However, our study had a number of limitations. Firstly, it was conducted over a very short period, during weather conditions that were warm and well suited to outdoor dining. About one-third of the restaurants observed had outdoor areas where smokers could eat and smoke. It will be important to reassess compliance with the legislation in the longer term and during colder months. Secondly, the selection of restaurants was not strictly random, but related to proximity to the homes of the volunteer observers. Selection from restaurant guides also meant that very low price cafes were not included, so our results can not be generalised to all commercial indoor eating venues. Thirdly, the reported perceptions of table waiters about customer reaction to smoke-free dining, although consistent with population-based surveys of dining preference,3,11-14 are unvalidated, as are their estimates of the volume of patronage. Proper validation of the impact of smoke-free legislation on restaurant and hotel trade would require the collection of longitudinal tax receipt data. In the United States, when tax receipts have been used as outcome measures, smoke-free laws have been shown to either have no effect on sales or to generate slightly increased patronage.5-8 Acknowledgements NSW Health funded the study. The New South Wales Cancer Council and Quit Victoria assisted in administering the grant. References Egan C. Where there's smoke, there's ire. The Australian 2000; 20 Dec. Rowbotham J. Restaurant, pub smoking ban stalled. Sydney Morning Herald 1999; 4 Feb: 6. Stevenson S. AHA to oppose new smoking ban moves. The Mercury (Hobart) 2000; 29 July: 9. Walsh RA, Paul CL, Tzelepis F. Overwhelming support for smoking bans. Aust N Z J Public Health 2000; 24: 640-641. Glantz SA. Effect of smoke-free bar law on bar revenues in California. Tob Control 2000; 9: 111-112. Glantz SA, Smith LRA. The effect of ordinances requiring smokefree restaurants on restaurant sales. Am J Public Health 1994; 84: 1081-1085. Chapman S, Lee NM, Monaem A. Smoking in restaurants [letter]. Med J Aust 1998; 168: 637. Glantz S, Charlesworth A. Tourism and hotel revenues before and after passage of smoke-free restaurant ordinances. JAMA 1999; 281: 1911-1918. Parliament of New South Wales. 52nd Parliament. Smoke-Free Environment Bill (Legislative Council). Hansard, 29 June 2000, p 7809. Government of Victoria. Department of Human Services, Public Health Division. Smoke free dining legislation — what does this mean for restaurants and cafes? Information Bulletin 1, May 2000. Available at <http://www.tobaccoreforms.vic.gov.au/cafes1.htm>. Jones A. Radio 2UE (Sydney) 30 Aug 2000; 06:41 am. Trotter L. Environmental tobacco smoke: surveys of restaurant patrons and hospitality industry personnel. Quit Evaluation Studies Vol. 9, Chapter 3. Available at: <http://www.quit.org.au/QE9/QE9Home.html>. Accessed 4 April 2001. Schofield MJ, Considine R, Boyle CA, Sanson-Fisher R. Smoking control in restaurants — the effectiveness of self-regulation in Australia. Am J Public Health 1993; 83: 1284-1288. Borland R, Hill D. Public attitudes to smoke-free zones in restaurants: an update [letter]. Med J Aust 1991; 154: 292-293. Mullins R. A survey of patrons of Melbourne restaurants on the provision of smokefree dining. Melbourne: Centre for Behavioural Research in Cancer, May 1991. Roberts C, Algert C, Chey T, Capon A. Community attitudes to smoking in restaurants [letter]. Med J Aust 1992; 157: 210. Shiel A, Chapman S. The inertia of self-regulation: a game-theoretic approach to reducing passive smoking in restaurants. Soc Sci Med 2000; 51: 1111-1119. Public smoking: the problem. Available at: <http://www.tobaccoinstitute.com/getallimg.asp?DOCID=TIMN0014554/4565>. Accessed 4 April 2001. Chapman S, Borland R, Brownson R, et al. The impact of workplace smoking bans on declining cigarette consumption in Australia and the USA. Am J Public Health 1999; 89: 1018-1023. Jarvis MJ, Foulds J, Feyerabend C. Exposure to passive smoking among bar staff. Br J Addict 1992; 87: 111-113. Laufer D. (Philip Morris USA) Presentation transcript. Available at: <http://www.pmdocs.com/getallimg. asp?DOCID=2041183751/3790>, p 28. Accessed 4 April 2001. (Received 2 Feb, accepted 27 Mar, 2001) Authors' details Department of Public Health and Community Medicine, University of Sydney, Sydney, NSW. Simon Chapman, BA(Hons), PhD, Professor of Public Health and Community Medicine. VicHealth Centre for Tobacco Control, Carlton, VIC. Ron Borland, PhD, Director; Anita Lal, BEc, GradDipPsych, Research Officer. Reprints will not be available from the authors. Correspondence: Professor S Chapman, Department of Public Health and Community Medicine, University of Sydney, Sydney, NSW 2006. simoncAThealth.usyd.edu.au Make a comment 1: Proportion of restaurants in Sydney and Melbourne in which various smoking behaviours were observed, correlated with individual restaurant policies on smoking Sydney Melbourne Observation Mandated ban (n = 78) Banned totally (n = 12) Banned in eating areas (n = 6) Partial bans* (n = 24) No restric- tions (n = 39) No smoking 100% 75% 50% 33% 21% Smoking, but not in eating areas 0 8% 50% 13% 8% Smoking in eating areas 0 0 0 50% 69% Ashtrays on tables 4% 0 0 38% 51% * Partial bans = arbitrary smoking and non-smoking sections. Back to text 2: Responses of restaurant staff in Sydney and Melbourne to questions about banning smoking in restaurants. Data are proportion of restaurants experiencing or anticipating problems/advantages by type of smoking ban Sydney Melbourne Observation/opinion Experience after introducing mandated ban (n = 78) Experience after introducing voluntary ban (n = 25) Expectations in restaurants without voluntary ban (n = 56) Patrons lighting up 31% 44% 49% Smokers complaining or "making a scene" 33% 56% 48% Smokers refusing to stop smoking when asked 6% 0 22% Smokers going elsewhere 21% 54% 51% Patrons saying they "won't be back" 4% 36% 24% Patrons stubbing out cigarettes on floor 0 12% 11% Patronage over past week Increased Normal Decreased 14% 76% 9% 24% 68% 8% 17% 72% 11% Increased number of non-smoker patrons qna 33% 49% Positive comments from patrons on ban None Once or twice Several times Many times Can't say 21% 23% 28% 28% 0 16% 16% 28% 36% 4% qna qna qna qna qna Staff support for bans Strong support Moderate support Neutral Against bans 49% 32% 14% 5% 76% 16% 4% 4% 41% 29% 18% 13% qna = question not asked. Back to text

Simon Chapman · Ron Borland · Anita Lal

Notable Cases Caffeine-induced cardiac arrhythmia: an unrecognised danger of healthfood products We describe a 25-year-old woman with pre-existing mitral valve prolapse who developed intractable ventricular fibrillation after consuming a "natural energy" guarana health drink containing a high concentration of caffeine. This case highlights the need for adequate labelling and regulation of such products. Marianne E Cannon, Clive T Cooke and James S McCarthy MJA 2001; 174: 520-521 Clinical record - Discussion - References - Authors' details - - More articles on Toxicology While caffeine is widely used in most Western societies, caffeine toxicity is rare, and certainly rarely diagnosed. Case reports of death from caffeine toxicity number approximately 20 in the medical literature.1-3 We report a case of sudden death in a young woman which was associated with ingestion of caffeine present in guarana, a widely available health supplement marketed as a "natural" source of energy, and targeted to young people in advertising. Clinical record A 25-year-old woman had been working at a bar where she was seen to collapse. Police at the scene started cardiopulmonary resuscitation shortly afterwards. An ambulance arrived several minutes later and she was defibrillated according to ambulance protocol, with a total of 12 defibrillation attempts. She arrived at the emergency department in ventricular fibrillation and was resuscitated according to advanced cardiac life support guidelines for a further 20 minutes. At no stage did she regain a spontaneous cardiac output. Further history (which became available later) indicated that the patient had been diagnosed with mitral valve prolapse. She had been referred to a cardiologist because of palpitations. He had raised the issue of caffeine intake, and she had agreed to limit this to a cup of tea daily. Her previously recorded resting electrocardiogram was normal, with no evidence of QT prolongation. On the day of her death she had been given a 55 mL squirt bottle of "Race 2005 Energy Blast with Guarana and Ginseng", which she had nearly emptied. Other staff working at the bar had also been given bottles of Race 2005. She had not consumed other caffeine-containing substances. At autopsy, the patient was found to have sclerosis and myxoid change of the mitral valve leaflets. A toxicological screen for all common prescribed and non-prescribed drugs (including opiates, cannabis, amphetamines, and cocaine metabolites) was performed. Assay methods included gas chromatography and mass specrometry (GC-MS) and immunoassay. Caffeine was detected by GC-MS; the toxicology screen was negative for other substances. High pressure liquid chromatography revealed a caffeine concentration of 19 mg/L (non-preserved) in aortic blood. The caffeine concentration in the bottle of Race 2005 Energy Blast was assayed at the Chemistry Centre of Western Australia and yielded a level of 10 g/L, which is more than 60 times the concentration of caffeine in cola beverages (see Box).4 A similar bottle was later shown to have a caffeine concentration of 19 g/L. Discussion The wide availability and toxicology of "natural" tonics and remedies and the regulation of such substances has been the subject of vigorous debate in both the lay press and the medical literature.6-10Guarana is produced from the seeds of the guarana plant (Paullina cupana), a creeping Amazonian shrub. The seeds are black "like an eye" and contain 3.6%-5.8% caffeine.11 The substance is named after the Guarani Amazonian tribe, who originally used the seeds to brew a drink.11 In Brazil, guarana is commonly used as an astringent, a flavouring, and as a stimulant, and it is increasingly being used in similar products internationally. Such products are widely available at pharmacies and natural food outlets, and are marketed as natural sources of energy. Caffeine is a natural alkaloid methylxanthine. Ninety-nine per cent is absorbed after oral ingestion, the blood concentration peaks 1-1.5 hours after ingestion, and its half-life in adults is 3-6 hours. Caffeine is metabolised by the P450 hepatic enzyme system.12 The pharmacodynamic profile of caffeine is similar to that of theophylline, another methylxanthine, in that it inhibits the adenosine receptor and acts as a phosphodiesterase inhibitor.12 At high serum levels, enzyme saturation occurs, and elimination follows zero-order kinetics, with constant elimination regardless of serum level.12 Caffeine increases intracellular calcium concentrations, causes noradrenaline release and sensitises dopamine receptors. The pharmacological effects of caffeine include central nervous system (CNS) and cardiac stimulation, as well as coronary vasodilatation. It relaxes smooth muscle, stimulates skeletal muscle, has a weak diuretic action, and its metabolic effects include hypoglycaemia. While toxicity generally occurs at serum levels over 25 mg/L, the correlation between concentration and clinical effects is poor.4 The measured concentration of 19 mg/L in our patient may reflect postmortem changes in drug level, or an enhanced sensitivity to caffeine toxicity in this patient. The measured concentration is the equivalent of what would result from drinking approximately 15-20 cups of coffee.4 The toxic effects of caffeine include vomiting and abdominal pain followed by CNS symptoms, including agitation, altered conscious state, rigidity and seizures.4 Cardiovascular effects include supraventricular and ventricular tachyarrhythmias, and significant metabolic disturbances may occur, including hypokalaemia and hyperglycaemia. Suggested management for caffeine toxicity, in addition to general supportive measures, includes multidose charcoal therapy, and extracorporeal elimination by charcoal haemoperfusion.13 One proposed mechanism for the seizures and cardiac arrhythmias is the blockade of adenosine receptors.12 Thus, there is a theoretical rationale for the use of adenosine in this setting, particularly if seizures are refractory to benzodiazepine agents.14 Likewise, as there is adrenergic stimulation, parenteral -blocker therapy (propranolol or esmolol) should be considered for treating ventricular arrhythmias.15 Most deaths associated with caffeine intoxication have occurred after overdose with diet pills and stimulants, and most have occurred in young patients without known underlying heart disease or variant of normal, such as mitral valve prolapse. At least two case reports document sudden death in patients who "walked" into an emergency department.14 After the death of our patient, the Western Australian Coroner recommended that Race 2005 Energy Blast be removed from the local market, and the product was recalled nationally by letter to distributors in August 1999. However, many other products containing guarana (eg, "energy" tablets containing 35 mg of caffeine each, as well as a variety of energy drinks) remain available Australia-wide. With the growing use of guarana-based products containing high levels of caffeine, there is an obvious potential for lethal overdose. Our patient had a relatively common cardiac abnormality, present in 2.4% of the population.16 While mitral valve prolapse has been associated with sudden death, the risk of this is low.17 Malignant arrhythmias may occur in the absence of cardiac stimulants, but the association of arrhythmia with drugs that increase cardiac irritability is well known. The role of caffeine in this woman's death is supported by her history of palpitations associated with caffeine ingestion, a history that had prompted her cardiologist to advise her to limit her caffeine intake. In this patient's case the Coroner found that the high level of caffeine was associated with the development of an intractable arrhythmia. This case highlights the need for more careful regulation of "natural" products, including warnings for patients with underlying health problems, and clear labelling to document the presence of any constituents with potentially toxic effects. It also shows the need for medical practitioners to be familiar with the more widely used "natural"-remedy substances, and their toxicological profile. Acknowledgements: We thank Dr Julian Stella and Dr Peter Sprivulis for their assistance in the preparation of this manuscript. Disclosure: We did not receive any financial or other support for this work, and have no financial or professional relationships that may pose a conflict of interest. References Garriott JC, Simmons LM, Poklis A, Mackell MA. Five cases of fatal overdose from caffeine-containing "look-alike" drugs. J Anal Toxicol 1985; 9: 141-143. Mrvos RM, Reilly PE, Dean BS, Krenzelok EP. Massive caffeine ingestion resulting in death. Vet Hum Toxicol 1989; 31: 571-572. Walsh I, Wasserman GS, Mestad P, Lanman RC. Near-fatal caffeine intoxication treated with peritoneal dialysis. Pediatr Emerg Care 1987; 3: 244-249. Lewin NA. Caffeine. In: Goldfrank LR, editor. Goldfrank's toxicologic emergencies. 6th ed. Stamford: Appleton & Lange, 1998: 555-562. Abbott PJ. Caffeine: a toxicological overview. Med J Aust 1986; 145: 518-521. Angell M, Kassirer JP. Alternative medicine — the risks of untested and unregulated remedies. N Engl J Med 1998; 339: 839-841. De Smet PA. Should herbal medicine-like products be licensed as medicines? BMJ 1995; 310: 1023-1024. Ernst E. Harmless herbs? A review of the recent literature. Am J Med 1998; 104: 170-178. Shaw D. Risks or remedies? Safety aspects of herbal remedies in the UK. J R Soc Med 1998; 91: 294-296. Kessler DA. Cancer and herbs. N Engl J Med 2000; 342: 1742-1743. Duke JA. CRC Handbook of medicinal herbs. Boca Raton: CRC Press, 1985. Serafin WE. Drugs used in the treatment of asthma. In: Hardman JG, Limbird LE, editors. Goodman & Gilman's the pharmacological basis of therapeutics. 9th ed. New York: McGraw-Hill, 1996. Nagesh RV, Murphy KA. Caffeine poisoning treated by hemoperfusion. Am J Kidney Dis 1988; 4: 316-318. Shum S, Seale C, Hathaway D, et al. Acute caffeine ingestion fatalities: management issues. Vet Hum Toxicol 1997; 39: 228-230. Pentel P. Toxicity of over-the-counter stimulants. JAMA 1984; 252: 1898-1903. Freed LA, Levy D, Levine RA, et al. Prevalence and clinical outcome of mitral-valve prolapse. N Engl J Med 1999; 341: 1-7. Duren DR, Becker AE, Dunning AJ. Long-term follow-up of idiopathic mitral valve prolapse in 300 patients: a prospective study. J Am Coll Cardiol 1988; 11: 42-47. (Received 5 Oct 2000, accepted 13 Mar 2001) Authors' details Department of Emergency Medicine, Fremantle Hospital, Fremantle, WA. Marianne E Cannon, MB BS, FACEM, Emergency Medicine Staff Specialist. PathCentre, QEII Medical Centre, Nedlands, WA. Clive T Cooke, BMedSci, FRCPA, Forensic Pathologist. University of Western Australia, Department of Medicine, Fremantle Hospital, Fremantle, WA. James S McCarthy, MB BS, FRACP, Senior Lecturer. Reprints will not be available from the authors. Correspondence: Dr M E Cannon, Department of Emergency Medicine, Fremantle Hospital, Fremantle, WA 6959. mariannecannonAThotmail.com Make a comment Caffeine content of various food, drinks and medicines Substance Defined dose* Concentration Tea 40mg per cup (approx.) 0.16g/L Coffee 100mg per cup (approx.) 0.4-1.6g/L Chocolate (30g) 4mg Cola 0.15g/L Diet pills 75-200mg Race 2005 (30mL "dose") 300-570mg 10-19g/L *After Lewin4 and Abbott.5 Back to text

Marianne E Cannon · Clive T Cooke · James S McCarthy