We declare no competing interests.
Michel Drancourt, *Didier Raoult [email protected]
Aix Marseille Université, Unité de Recherche sur les Maladies Infectieuses et Tropicales Émergentes (URMITE), UMR63 CNRS 7278, IRD 198, INSERM 1095, 13005 Marseille, France 1
Wagner DM, Klunk J, Harbeck M, et al. Yersinia pestis and the Plague of Justinian 541–543 AD: a genomic analysis. Lancet Infect Dis 2014; 14: 319–26. Drancourt M, Aboudharam G, Signoli M, et al. Detection of 400-year-old Yersinia pestis DNA in human dental pulp: an approach to the diagnosis of ancient septicemia. Proc Natl Acad Sci USA 1998; 95: 12637–40. Gilbert MTP. Yersinia pestis: one pandemic, two pandemics, three pandemics, more? Lancet Infect Dis 2014; 14: 264–65. Gilbert MT, Cuccui J, White W, et al. Absence of Yersinia pestis-speciﬁc DNA in human teeth from ﬁve European excavations of putative plague victims. Microbiology 2004; 150: 341–54. Raoult D. Was the Black Death yersinial plague? Lancet Infect Dis 2003; 3: 328.
Authors’ reply Monica Green and colleagues take issue with three claims made in our Article1 because they say they are unproven. We suggested that Yersinia pestis might have travelled overland from China to Europe to cause the ﬁrst plague pandemic, spread from Europe or north Africa to central Africa during the second plague pandemic leading to the establishment of the 1.ANT group, and spread back to Asia from Europe or north Africa giving rise to the 1.IN populations in China. We described these ideas as a viable explanation for the phylogeographic patterns that we noted and a “hypothetical scenario for the geographic spread of Yersinia pestis”. Green and colleagues might have had issue with our statement that “Historical records are consistent with this scenario…” possibly because they took this statement out of context or misunderstood our intent. Our intent was to point out that historical information, as recorded by Robert Pollitzer,2 describes the spread of plague from Yunnan province, China, starting around 1855. This spread of plague is widely considered to be the start of the third plague pandemic, which involved the eventual worldwide spread of the www.thelancet.com/infection Vol 14 October 2014
1.ORI group.3–5 We pointed out that the 1.IN3 group is also currently found in Yunnan province,4 where the third pandemic is thought to have originated. In addition to the co-occurrence of these groups in Yunnan province, we noted,1 as have others,4,5 that the most closely related group to the 1.ORI group is the 1.IN3 group (which is again the group associated with the third pandemic). Finally, we stated that Yujun Cui and colleagues4 reported that some 1.IN3 strains from Yunnan province possess the Orientalis phenotype, which is conserved in all known 1.ORI strains, but not identiﬁed in any other groups within Y pestis except these few 1.IN3 strains from Yunnan province. In view of all of these patterns, we suggested that the 1.IN populations could have given rise to the 1.ORI group. Science leads to advances by putting forth hypotheses that can be tested by subsequent studies. We never suggested that our ideas that are disputed by Green and colleagues were proven; rather, they are hypotheses that can be tested with additional sampling of ancient and contemporary Y pestis samples. Indeed, we are actively pursuing Y pestis samples that can be used to rigorously test these hypotheses and we encourage other researchers to do the same. We are confused by the comments made by Michel Drancourt and Didier Raoult who claim to have shown the microbial cause of the Justinianic plague in their 1998 paper, because that paper reports the potential presence of Y pestis in samples from 16th century France, which are 1000 years more recent than those described by us1 and from an entirely different pandemic. In a later study6 they claimed that the contemporary Orientalis biovar of Y pestis was found in teeth dating back to the Justinianic plague, which is at odds with our discovery of a unique lineage that is clearly distinct from lineages currently circulating in human and animal populations, as well as ﬁnding by others4,5 that the Orientalis biovar probably arose within the past
200–210 years and, therefore, is much too recent to have been associated with the Justinianic plague. The methods they used in this later study have also been questioned.7 Finally, despite claims to the contrary, the work of Drancourt and Raoult is not the ﬁrst use of DNA from dental pulp.8 We declare no competing interests.
David M Wagner, Paul S Keim, Holger C Scholz, Edward C Holmes, *Hendrik Poinar [email protected]
Department of Biological Sciences and Center for Microbial Genetics and Genomics, Northern Arizona University, Flagstaﬀ, AZ, USA (DMW, PSK); Bundeswehr Institute for Microbiology, Munich, Germany (HCS); Marie Bashir Institute for Infectious Diseases and Biosecurity, University of Sydney, Sydney, NSW, Australia (ECH); and McMaster Ancient DNA Centre, Department of Anthropology, McMaster University, Hamilton, ON L8S 4L8, Canada (HP) 1
Wagner DM, Klunk J, Harbeck M, et al. Yersinia pestis and the Plague of Justinian 541–543 AD: a genomic analysis. Lancet Infect Dis 2014; 14: 319–26. Pollitzer R. Plague studies. 1. A summary of the history and survey of the present distribution of the disease. Bull World Health Organ 1951; 4: 475–533. Achtman M, Morelli G, Zhu P, et al. Microevolution and history of the plague bacillus, Yersinia pestis. Proc Natl Acad Sci USA 2004; 101: 17837–42. Cui Y, Yu C, Yan Y, et al. Historical variations in mutation rate in an epidemic pathogen, Yersinia pestis. Proc Natl Acad Sci USA 2013; 110: 577–82. Morelli G, Song Y, Mazzoni CJ, et al. Yersinia pestis genome sequencing identiﬁes patterns of global phylogenetic diversity. Nat Genet 2010; 42: 1140–43. Drancourt M, Roux V, Dang LV, et al. Genotyping, Orientalis-like Yersinia pestis, and plague pandemics. Emerg Infect Dis 2004; 10: 1585–92. Vergnaud G. Yersinia pestis genotyping. Emerg Infect Dis 2005; 11: 1317–18. Potsch L, Meyer U, Rotschild S, Schneider P, Rittner CH. Application of DNA techniques for identiﬁcation using human dental pulp as a source of DNA. Intl J Leg Med 1992; 105: 139–43.
Antibiotics for acute sore throat We read Paul Little and colleagues’ report 1 of their cohort study of antibiotic prescription strategies for acute sore throat with great interest. The authors conclude that in most patients, antibiotics are not needed, but that a delayed prescription 919
strategy might be as beneficial as immediate prescription. However, we do not support a delayed prescription strategy, for the following reasons. Guidelines from the UK National Institute for Health and Care Excellence and the Scottish Intercollegiate Guidelines Network state that “antibiotics should not be used to secure symptomatic relief in sore throat”. Moreover, delayed prescription shifts responsibility to patients.2 Additionally, in a previous study by Little and colleagues, 3 quinsy occurred in 30 (0·5%) of 5932 patients who were given antibiotics compared with 11 (0·2%) of 4974 who were not given antibiotics. This ﬁnding needs more explanation and could be used to support the argument against intake of antibiotics. We agree with the Comment2 by Benedikt Huttner that Little and colleagues’ cohort study 1 is prone to confounding—eg, the types of antibiotics prescribed are unknown. Furthermore, information about what proportion of patients actually took their prescribed antibiotics (delayed or immediate prescriptions) is not provided. However, most importantly, most studies into antibiotics for sore throat have reported that the greatest improvement of symptoms occurs 3 or 4 days after treatment begins (for the 20% patients with a proven group A β-haemolytic streptococcal pharyngitis). 4,5 Thus, delaying antibiotics by 48 h or more would overshoot this period.4 We agree with diﬀerent European guidelines for acute sore throat, which recommend antibiotics only for very ill or immunocompromised patients (about 5% of patients with acute sore throat, which is far fewer than would receive antibiotics in a delayed prescription strategy). 5 To achieve the least amount of antibiotic use—in the absence of risk to the patient—doctors should use a noantibiotics strategy with advice to 920
return to the clinic if symptoms do not resolve.4,5 Paracetamol (or nonsteroidal anti-inflammatory drugs) taken every 4 h for an extra 1–2 days could replace unnecessary antibiotic use in most patients and might reduce antibiotic resistance. The best strategy to reduce antibiotic use for acute sore throat is to motivate doctors not to prescribe antibiotics, instead of putting this responsibility on the shoulders of patients. We declare no competing interests.
*Jan Matthys, Marc De Meyere [email protected]
University of Ghent, De Pintelaan 185, Ghent 9000, Belgium 1
Little P, Stuart B, Hobbs FD, et al, for the DESCARTE investigators. Antibiotic prescription strategies for acute sore throat: a prospective observational cohort study. Lancet Infect Dis 2014; 14: 213–19. Huttner B. Antibiotic prescription for sore throat or the legacy of Mr X2. Lancet Infect Dis 2014; 14: 177–78. Little P, Stuart B, Hobbs FD, et al, for the DESCARTE investigators. Predictors of suppurative complications for acute sore throat in primary care: prospective clinical cohort study. BMJ 2013; 347: f6867. Spurling GK, Del Mar CB, Dooley L, Foxlee R, Farley R. Delayed antibiotics for respiratory infections. Cochrane Database Syst Rev 2013; 4: CD004417. Matthys J, De Meyere M, van Driel ML, De Sutter A. Diﬀerences among international pharyngitis guidelines: not just academic. Ann Fam Med 2007; 5: 436–43.
Author’s reply Jan Matthys and Marc De Meyere’s have several criticisms of our paper. First, with respect to their comment that delayed prescription shifts the responsibility to patients, one of the reasons delayed prescription is likely to be effective is because it shifts some responsibility to patients, who are also satisfied with the approach.1,2 It reduces both patients’ belief in antibiotics and reconsultations,1,2 and reconsultations often result in an antibiotic prescription. The fact that in our previous study quinsy occurred in numerically more patients who were given antibiotics than those who were not is not unusual: the crude figures do not
take into account confounding by indication, which our current paper does. We agree that our study was prone to confounding, but when the factors that predict prescription are well documented (as in our current study), propensity scores can usefully correct for confounding by indication. Residual confounding is always possible, but the direction and magnitude of the change in estimates using propensity scores would suggest that the effect of residual confounding, if anything, would be to underestimate the beneﬁt of delayed prescribing. The assertion that the time of greatest benefit for symptoms is 3–4 days after treatment does not seem to be supported by the results of a Cochrane review,3 which show that the greatest risk reduction in absolute and relative terms occurs at day 7 not day 3, and an even greater diﬀerence in risk ratios at day 7 for those with, compared with those without, streptococcal infection. With respect to guidelines’ suggestions that antibiotics should only be used in very ill or immunocompromised patients, we agree that antibiotics are mostly not necessary for symptom control, but the definition of “very ill” is subjective. Data suggest that when a doctor is concerned about the risk of complications or severe symptoms and is considering prescribing antibiotics, delayed antibiotics can be useful, and can reduce reconsultations and complications, compared with a no-prescription strategy. This accords with the UK National Institute for Health and Care Excellence (NICE), 4 which supports delayed prescription as a cost-effective strategy, and NICE guidance has the dual advantages of both a systematic scientiﬁc literature review and economic modelling to inform policy, unlike many guidelines. I declare no competing interests.
www.thelancet.com/infection Vol 14 October 2014