- Email: [email protected]
Strategies for coronary stenting
COMMENTARY
and colleagues to determine viral load is more quantitative than that in commercially available tests, a specific viral load that defines a high risk has not yet been determined.5,6 Josefsson and co-workers found that the positive-predictive value of a measurement of HPV 16 DNA was low and unlikely to be useful as a single predictor of CIS risk except in the high-load group among whom the positive-predictive value was 48%. The researchers suggest that tests to detect other oncogenic HPV types and carrying out the tests on fresh specimens may improve the positivepredictive value. As Ylitalo and colleagues suggest, quantification of viral load might be added to cytology to reduce the number of women referred for treatment of mildly abnormal Pap smears. Thus the viral load of HPV 16 may be used to identify women with normal and mildly abnormal smears who are at risk for high-grade disease. If these suggestions are carried out, several questions arise. What should be done for these ostensibly high-risk women, once they are identified? Since there is currently no vaccine or treatment for latent HPV infections, are these women candidates for chemoprevention trials? How should the women be followed in the absence of documented SIL? Considering accessibility, cost, and positive-predictive value, to what population should viral tests be applied to make the greatest impact? Finally, whether any of these new tests will be a substantial improvement to the accessibility, cost, and positive-predictive value of the Pap smear, has yet to be seen. Carolyn Johnston Department of Obstetrics and Gynecology, University of Michigan Hospital, Ann Arbor, MI 48109-0276, USA 1
Liaw KL, Glass AG, Manos MM, et al. Detection of human papillomavirus DNA in cytologically normal women and subsequent cervical squamous intraepithelial lesions. J Natl Cancer Inst 1999; 91: 954–60. 2 Koutsky LA, Holmes KK, Critchlow CW, et al. A cohort study of the risk of cervical intraepithelial neoplasia grade 2 or 3 in relation to papillomavirus infection. N Engl J Med 1992; 327: 1272–78. 3 Campion MJ, McCance DJ, Cuzick J, et al. Progressive potential of mild cervical atypia: prospective cytological colposcopic, and virological study. Lancet 1986; ii: 237–40. 4 Ho GY, Burk RD, Klein S, et al. Persistent genital human papillomavirus infection as a risk factor for persistent cervical dysplasia. J Natl Cancer Inst 1995; 87: 1365–71. 5 Cox JT, Lorincz AT, Schiffman MH, et al. Human papillomavirus testing by hybrid capture appears to be useful in triaging women with a cytologic diagnosis of atypical squamous cells of undetermined significance. Am J Obstet Gynecol 1995; 172: 946–54. 6 Swan DC, Tucker RA, Tortolero-Luna G, et al. Human papillomavirus (HPV) DNA copy number is dependent on grade of cervical disease and HPV type. J Clin Microbiol 1999; 37: 1030–34. 7 Manos MM, Kinney WK, Hurley LB, et al. Identifying women with cervical neoplasia using human papillomavirus DNA testing for equivocal Papanicolaou results. JAMA 1999; 281: 1605–10. 8 Schiffman M, Herrero R, Hildesheim A, et al. HPV DNA testing in cervical cancer screening. Results from women in a high-risk province of Costa Rica. JAMA 2000; 283: 87–93. 9 Wright TC, Denny L, Kuhn L, et al. HPV DNA testing of selfcollected vaginal samples compared with cytologic screening to detect cervical cancer. JAMA 2000; 283: 81–86. 10 Hillemanns P, Kimmig R, Huttemann U, et al. Screening for cervical neoplasia by self-assessment for human papillomavirus DNA. Lancet 1999; 354: 1970. 11 Wright TC, Sun XW, Koulos J. Comparison of management algorithms for the evaluation of women with low-grade cytologic abnormalities. Obstet Gynecol 1995; 85: 202–10.
2180
Strategies for coronary stenting See page 2199
Most percutaneous coronary revascularisation procedures involve remodelling of the vessel lumen with balloon dilatation. Although stent implantation as an adjunct to balloon angioplasty has proven valuable for several types of devices and for a range of clinical indications, the use of stents in other settings is still uncertain. For example, more data are needed to support the current high rate of stent usage in vessels of small or large calibre (diameter <2·5 mm or >3·5mm). There is also uncertainty about the most effective strategy for clinical application.The OPUS 1 trial reported in this week’s issue of The Lancet makes an important contribution towards resolving this latter uncertainty and supports the recent recommendations for the use of coronary artery stents in the treatment of ischaemic heart disease issued by the UK National Institute for Clinical Excellence (available at www.nice.org.uk). Several approaches to stent use have been studied and debate has been hampered by confusion over terminology. A brief review of the titles of the references of the OPUS 1 report illustrates this difficulty. Perhaps the following descriptive scheme may be useful. “Primary stenting” would describe strategies that involve the predetermined intention to treat a lesion with stent implantation. Although this approach has also been described as “routine stenting” (as in OPUS 1), “elective stenting” or “systematic stenting”, these terms are less favourable since they may be confused with other, more widely applied descriptions of clinical activity. Case and lesion complexity, comorbidity, and complications may render many primary stent implantations far from routine. Similarly, an emergency clinical presentation might undermine the concept of “elective”. The term “direct stenting” should be reserved for the subset of primary stent implantation procedures where deployment is carried out without initial balloon predilatation. This new, but popular, technique is suitable for non-occlusive lesions and is currently being evaluated in ongoing studies. The imprecise description of “direct implantation” as “primary stenting” (of which it is a form), is the most common source of confusion in interventional debate. All other forms of stent use are “provisional”. This term assumes that an initial attempt is made to treat the target lesion with alternative devices—generally an angioplasty balloon—and the stent is deployed if the result of the primary treatment device is suboptimum. This approach has also been termed “conditional stenting”. When stent deployment is required to manage abrupt or threatened vessel closure (a small but important subgroup of provisional implants) then this constitutes “bailout stenting”. The methodology of the placebo-controlled, randomised drug trial does not readily translate to a comparison of two strategies for a complex interventional procedure. The OPUS 1 investigators are to be congratulated for their pragmatic trial design that attempts to mirror routine clinical practice. Such a design should allow better generalisation of their findings. However, the study does have some important limitations, knowledge of which may guide the design and conduct of future ventures. OPUS 1 was stopped early due to a decline in recruitment presumably related to the increasing popularity of stenting and a reluctance, on behalf of both
THE LANCET • Vol 355 • June 24, 2000
For personal use only. Not to be reproduced without permission of The Lancet.
COMMENTARY
clinicians and patients, to accept randomisation to a provisional stent strategy. Earlier initiation of the study might have avoided this difficulty but would have had to be balanced against a formal assessment of a new procedure too early in its clinical development. The double-open design meant that both patients and doctors were aware of the treatment allocation. The lack of masking may assume particular importance when key outcome measures involve perceptions or value judgments by individuals, for example, the need for target-vessel revascularisation. The treatment of in-stent restenosis can be difficult and is a less appealing prospect than primary stent implantation for balloon-angioplasty restenosis. Most of the participants in the trial had single-vessel coronary disease, and it is possible that recurrent symptoms in a proportion of the primary stent group could have been managed with intensification of medical therapy, thus limiting the number of additional revascularisation procedures. Certainly it seems odd that the patientreported quality of life and functional status were similar in the two groups despite the significantly lower frequency of the composite endpoint in the primary stent group. Some studies elect to perform routine coronary angiography at follow-up, but this approach is not suitable for pragmatic strategy trials. Centralised computer analysis of angiographic records provides precise information about vessel geometry that can be important in understanding the mechanisms of restenosis, but it is difficult to relate these data to clinical outcomes. Furthermore, routine diagnostic angiography may drive additional revascularisation procedures that are not clinically indicated.1 It may be useful if study documentation demanded a more complete and structured reporting of clinical assessment throughout follow-up. Reported symptoms, details of medical therapy, and objective measures of functional capacity could be recorded, ideally by observers unaware of the treatment allocation. The two treatment strategies might involve different periods of specific risk for outcome measures. Follow-up periods should be extended to allow these outcome measures to accrue. These measures would not eliminate the difficulties of potential bias in a complex clinical area but would afford an opportunity to examine trial results for more predictable effects. R H Stables Cardiothoracic Centre and Royal Liverpool University Hospital, Liverpool L7 8XP, UK 1
Ruygrok PN, Melkert R, Morel MA, et al. Does angiography six months after coronary intervention influence management and outcome? Benestent II Investigators. J Am Coll Cardiol 1999; 34: 1507–11.
Insulin resistance, iron, and the liver In a report in 1997,1 a French group led by Yves Deugnier described a new clinical-pathological syndrome of primary hepatic iron overload (HIO), in which the transferrin saturation was normal, the frequency of the HLA A3 genotype (which is linked to genetic haemochromatosis) was normal, and serum and liver iron concentrations were raised disproportionately to the increase in serum ferritin concentration. Among the 65 patients with this syndrome, 72% had a body-mass index (BMI) greater than 25 kg/m2, 65% were dyslipidaemic, and 43% had abnormal glucose tolerance—ie, there was a high frequency of some of the features of the insulin-resistance syndrome (IRS).2 In a
THE LANCET • Vol 355 • June 24, 2000
subsequent study of 161 patients,3 the same group of investigators broadened the scope of unexplained HIO by including patients with increased transferrin saturation or non-alcoholic steato-hepatitis (NASH—ie, steatosis accompanied by lobular inflammation, sinusoidal fibrosis, and hepatocyte degeneration/necrosis). No patients were homozygous for Cys282Tyr but a higher than expected proportion of individuals were compound heterozygotes for the two main point mutations (Cys282Tyr and His63Asp) of HFE (the gene associated with genetic haemochromatosis), a finding that is expected in a cohort selected on the basis of iron overload. Also, with this extended definition, the patients with the syndrome of unexplained HIO, mostly middle-aged men, had a high rate of IRS (94%), by the investigators’ definition. Syndromes are clusters of non-chance associations, and the components of a syndrome can generally be related to a common element. In IRS, central fat accumulation (truncal or abdominal/visceral), impaired glucose metabolism (or overt diabetes mellitus), raised blood pressure (or clinical hypertension), and dyslipidaemia (typically, higher serum triglyceride concentrations along with lower serum HDLcholesterol concentrations, and small, dense LDL particles) cluster with insulin resistance, as indicated by a low totalbody insulin-mediated glucose disposal on euglycaemic clamp studies. Since clamp studies are time-consuming and labour-intensive, fasting hyperinsulinaemia is the most commonly used surrogate for insulin resistance.4 Obesity— as defined by the BMI—is not an essential component of the IRS,2,4 but if it were a criterion it would increase the prevalence of the syndrome because of the frequent occurrence of insulin resistance in obese individuals.5 In the studies of primary HIO,1,3 or hepatic steatosis,6–9 the IRS has been defined in differing ways, or with too broad criteria.1,3 For example, the French group in their study of patients with HIO,3 used obesity as a criterion but did not report actual BMI values or any measure of fat distribution (waist circumference or ratio of waist-to-hip circumference), or plasma insulin concentrations. Nevertheless, the occurrence of HIO, hepatic steatosis, or both among patients with the IRS is probably higher than among the insulin-sensitive proportion of the population. In particular, middle-aged men without symptoms who are investigated for raised serum iron concentrations are likely to have an IRS-like phenotype. Conversely, it is probable that screening individuals with features of the IRS would identify many cases of non-genetic iron overload. Several issues need to be clarified, however, before a new syndrome—insulin-resistance-associated HIO or IRHIO, as proposed by the French group—is accepted as a distinct entity. First, the IRS should be identified by consistent criteria, among which hyperinsulinaemia, or better, insulin resistance on euglycaemic clamp, is essential.10 Second, the diagnosis of HIO is not easy and needs a simple and reliable method of confirmation.11 Third, IRHIO seems to distinctly prefer men, but the IRS is common among postmenopausal women: are insulin-resistant women protected from iron overload? Fourth, since the IRS includes links of variable strength, it is important to determine which component of the IRS best correlates with HIO. Visceral obesity seems to be the best candidate, since increasing evidence links visceral-fat accumulation with widespread lipid deposition in liver parenchyma.6,7 Finally, if the IRS and HIO are part of the same syndrome, they should be modified in parallel by manoeuvres that affect either feature. For example, in an overweight individual weight loss generally leads to a 2181
For personal use only. Not to be reproduced without permission of The Lancet.
and colleagues to determine viral load is more quantitative than that in commercially available tests, a specific viral load that defines a high risk has not yet been determined.5,6 Josefsson and co-workers found that the positive-predictive value of a measurement of HPV 16 DNA was low and unlikely to be useful as a single predictor of CIS risk except in the high-load group among whom the positive-predictive value was 48%. The researchers suggest that tests to detect other oncogenic HPV types and carrying out the tests on fresh specimens may improve the positivepredictive value. As Ylitalo and colleagues suggest, quantification of viral load might be added to cytology to reduce the number of women referred for treatment of mildly abnormal Pap smears. Thus the viral load of HPV 16 may be used to identify women with normal and mildly abnormal smears who are at risk for high-grade disease. If these suggestions are carried out, several questions arise. What should be done for these ostensibly high-risk women, once they are identified? Since there is currently no vaccine or treatment for latent HPV infections, are these women candidates for chemoprevention trials? How should the women be followed in the absence of documented SIL? Considering accessibility, cost, and positive-predictive value, to what population should viral tests be applied to make the greatest impact? Finally, whether any of these new tests will be a substantial improvement to the accessibility, cost, and positive-predictive value of the Pap smear, has yet to be seen. Carolyn Johnston Department of Obstetrics and Gynecology, University of Michigan Hospital, Ann Arbor, MI 48109-0276, USA 1
Liaw KL, Glass AG, Manos MM, et al. Detection of human papillomavirus DNA in cytologically normal women and subsequent cervical squamous intraepithelial lesions. J Natl Cancer Inst 1999; 91: 954–60. 2 Koutsky LA, Holmes KK, Critchlow CW, et al. A cohort study of the risk of cervical intraepithelial neoplasia grade 2 or 3 in relation to papillomavirus infection. N Engl J Med 1992; 327: 1272–78. 3 Campion MJ, McCance DJ, Cuzick J, et al. Progressive potential of mild cervical atypia: prospective cytological colposcopic, and virological study. Lancet 1986; ii: 237–40. 4 Ho GY, Burk RD, Klein S, et al. Persistent genital human papillomavirus infection as a risk factor for persistent cervical dysplasia. J Natl Cancer Inst 1995; 87: 1365–71. 5 Cox JT, Lorincz AT, Schiffman MH, et al. Human papillomavirus testing by hybrid capture appears to be useful in triaging women with a cytologic diagnosis of atypical squamous cells of undetermined significance. Am J Obstet Gynecol 1995; 172: 946–54. 6 Swan DC, Tucker RA, Tortolero-Luna G, et al. Human papillomavirus (HPV) DNA copy number is dependent on grade of cervical disease and HPV type. J Clin Microbiol 1999; 37: 1030–34. 7 Manos MM, Kinney WK, Hurley LB, et al. Identifying women with cervical neoplasia using human papillomavirus DNA testing for equivocal Papanicolaou results. JAMA 1999; 281: 1605–10. 8 Schiffman M, Herrero R, Hildesheim A, et al. HPV DNA testing in cervical cancer screening. Results from women in a high-risk province of Costa Rica. JAMA 2000; 283: 87–93. 9 Wright TC, Denny L, Kuhn L, et al. HPV DNA testing of selfcollected vaginal samples compared with cytologic screening to detect cervical cancer. JAMA 2000; 283: 81–86. 10 Hillemanns P, Kimmig R, Huttemann U, et al. Screening for cervical neoplasia by self-assessment for human papillomavirus DNA. Lancet 1999; 354: 1970. 11 Wright TC, Sun XW, Koulos J. Comparison of management algorithms for the evaluation of women with low-grade cytologic abnormalities. Obstet Gynecol 1995; 85: 202–10.
2180
Strategies for coronary stenting See page 2199
Most percutaneous coronary revascularisation procedures involve remodelling of the vessel lumen with balloon dilatation. Although stent implantation as an adjunct to balloon angioplasty has proven valuable for several types of devices and for a range of clinical indications, the use of stents in other settings is still uncertain. For example, more data are needed to support the current high rate of stent usage in vessels of small or large calibre (diameter <2·5 mm or >3·5mm). There is also uncertainty about the most effective strategy for clinical application.The OPUS 1 trial reported in this week’s issue of The Lancet makes an important contribution towards resolving this latter uncertainty and supports the recent recommendations for the use of coronary artery stents in the treatment of ischaemic heart disease issued by the UK National Institute for Clinical Excellence (available at www.nice.org.uk). Several approaches to stent use have been studied and debate has been hampered by confusion over terminology. A brief review of the titles of the references of the OPUS 1 report illustrates this difficulty. Perhaps the following descriptive scheme may be useful. “Primary stenting” would describe strategies that involve the predetermined intention to treat a lesion with stent implantation. Although this approach has also been described as “routine stenting” (as in OPUS 1), “elective stenting” or “systematic stenting”, these terms are less favourable since they may be confused with other, more widely applied descriptions of clinical activity. Case and lesion complexity, comorbidity, and complications may render many primary stent implantations far from routine. Similarly, an emergency clinical presentation might undermine the concept of “elective”. The term “direct stenting” should be reserved for the subset of primary stent implantation procedures where deployment is carried out without initial balloon predilatation. This new, but popular, technique is suitable for non-occlusive lesions and is currently being evaluated in ongoing studies. The imprecise description of “direct implantation” as “primary stenting” (of which it is a form), is the most common source of confusion in interventional debate. All other forms of stent use are “provisional”. This term assumes that an initial attempt is made to treat the target lesion with alternative devices—generally an angioplasty balloon—and the stent is deployed if the result of the primary treatment device is suboptimum. This approach has also been termed “conditional stenting”. When stent deployment is required to manage abrupt or threatened vessel closure (a small but important subgroup of provisional implants) then this constitutes “bailout stenting”. The methodology of the placebo-controlled, randomised drug trial does not readily translate to a comparison of two strategies for a complex interventional procedure. The OPUS 1 investigators are to be congratulated for their pragmatic trial design that attempts to mirror routine clinical practice. Such a design should allow better generalisation of their findings. However, the study does have some important limitations, knowledge of which may guide the design and conduct of future ventures. OPUS 1 was stopped early due to a decline in recruitment presumably related to the increasing popularity of stenting and a reluctance, on behalf of both
THE LANCET • Vol 355 • June 24, 2000
For personal use only. Not to be reproduced without permission of The Lancet.
COMMENTARY
clinicians and patients, to accept randomisation to a provisional stent strategy. Earlier initiation of the study might have avoided this difficulty but would have had to be balanced against a formal assessment of a new procedure too early in its clinical development. The double-open design meant that both patients and doctors were aware of the treatment allocation. The lack of masking may assume particular importance when key outcome measures involve perceptions or value judgments by individuals, for example, the need for target-vessel revascularisation. The treatment of in-stent restenosis can be difficult and is a less appealing prospect than primary stent implantation for balloon-angioplasty restenosis. Most of the participants in the trial had single-vessel coronary disease, and it is possible that recurrent symptoms in a proportion of the primary stent group could have been managed with intensification of medical therapy, thus limiting the number of additional revascularisation procedures. Certainly it seems odd that the patientreported quality of life and functional status were similar in the two groups despite the significantly lower frequency of the composite endpoint in the primary stent group. Some studies elect to perform routine coronary angiography at follow-up, but this approach is not suitable for pragmatic strategy trials. Centralised computer analysis of angiographic records provides precise information about vessel geometry that can be important in understanding the mechanisms of restenosis, but it is difficult to relate these data to clinical outcomes. Furthermore, routine diagnostic angiography may drive additional revascularisation procedures that are not clinically indicated.1 It may be useful if study documentation demanded a more complete and structured reporting of clinical assessment throughout follow-up. Reported symptoms, details of medical therapy, and objective measures of functional capacity could be recorded, ideally by observers unaware of the treatment allocation. The two treatment strategies might involve different periods of specific risk for outcome measures. Follow-up periods should be extended to allow these outcome measures to accrue. These measures would not eliminate the difficulties of potential bias in a complex clinical area but would afford an opportunity to examine trial results for more predictable effects. R H Stables Cardiothoracic Centre and Royal Liverpool University Hospital, Liverpool L7 8XP, UK 1
Ruygrok PN, Melkert R, Morel MA, et al. Does angiography six months after coronary intervention influence management and outcome? Benestent II Investigators. J Am Coll Cardiol 1999; 34: 1507–11.
Insulin resistance, iron, and the liver In a report in 1997,1 a French group led by Yves Deugnier described a new clinical-pathological syndrome of primary hepatic iron overload (HIO), in which the transferrin saturation was normal, the frequency of the HLA A3 genotype (which is linked to genetic haemochromatosis) was normal, and serum and liver iron concentrations were raised disproportionately to the increase in serum ferritin concentration. Among the 65 patients with this syndrome, 72% had a body-mass index (BMI) greater than 25 kg/m2, 65% were dyslipidaemic, and 43% had abnormal glucose tolerance—ie, there was a high frequency of some of the features of the insulin-resistance syndrome (IRS).2 In a
THE LANCET • Vol 355 • June 24, 2000
subsequent study of 161 patients,3 the same group of investigators broadened the scope of unexplained HIO by including patients with increased transferrin saturation or non-alcoholic steato-hepatitis (NASH—ie, steatosis accompanied by lobular inflammation, sinusoidal fibrosis, and hepatocyte degeneration/necrosis). No patients were homozygous for Cys282Tyr but a higher than expected proportion of individuals were compound heterozygotes for the two main point mutations (Cys282Tyr and His63Asp) of HFE (the gene associated with genetic haemochromatosis), a finding that is expected in a cohort selected on the basis of iron overload. Also, with this extended definition, the patients with the syndrome of unexplained HIO, mostly middle-aged men, had a high rate of IRS (94%), by the investigators’ definition. Syndromes are clusters of non-chance associations, and the components of a syndrome can generally be related to a common element. In IRS, central fat accumulation (truncal or abdominal/visceral), impaired glucose metabolism (or overt diabetes mellitus), raised blood pressure (or clinical hypertension), and dyslipidaemia (typically, higher serum triglyceride concentrations along with lower serum HDLcholesterol concentrations, and small, dense LDL particles) cluster with insulin resistance, as indicated by a low totalbody insulin-mediated glucose disposal on euglycaemic clamp studies. Since clamp studies are time-consuming and labour-intensive, fasting hyperinsulinaemia is the most commonly used surrogate for insulin resistance.4 Obesity— as defined by the BMI—is not an essential component of the IRS,2,4 but if it were a criterion it would increase the prevalence of the syndrome because of the frequent occurrence of insulin resistance in obese individuals.5 In the studies of primary HIO,1,3 or hepatic steatosis,6–9 the IRS has been defined in differing ways, or with too broad criteria.1,3 For example, the French group in their study of patients with HIO,3 used obesity as a criterion but did not report actual BMI values or any measure of fat distribution (waist circumference or ratio of waist-to-hip circumference), or plasma insulin concentrations. Nevertheless, the occurrence of HIO, hepatic steatosis, or both among patients with the IRS is probably higher than among the insulin-sensitive proportion of the population. In particular, middle-aged men without symptoms who are investigated for raised serum iron concentrations are likely to have an IRS-like phenotype. Conversely, it is probable that screening individuals with features of the IRS would identify many cases of non-genetic iron overload. Several issues need to be clarified, however, before a new syndrome—insulin-resistance-associated HIO or IRHIO, as proposed by the French group—is accepted as a distinct entity. First, the IRS should be identified by consistent criteria, among which hyperinsulinaemia, or better, insulin resistance on euglycaemic clamp, is essential.10 Second, the diagnosis of HIO is not easy and needs a simple and reliable method of confirmation.11 Third, IRHIO seems to distinctly prefer men, but the IRS is common among postmenopausal women: are insulin-resistant women protected from iron overload? Fourth, since the IRS includes links of variable strength, it is important to determine which component of the IRS best correlates with HIO. Visceral obesity seems to be the best candidate, since increasing evidence links visceral-fat accumulation with widespread lipid deposition in liver parenchyma.6,7 Finally, if the IRS and HIO are part of the same syndrome, they should be modified in parallel by manoeuvres that affect either feature. For example, in an overweight individual weight loss generally leads to a 2181
For personal use only. Not to be reproduced without permission of The Lancet.