Skip directly to site content Skip directly to page options Skip directly to A-Z link Skip directly to A-Z link Skip directly to A-Z link
Volume 26, Number 10—October 2020

Deaths Associated with Pneumonic Plague, 1946–2017

Article Metrics
citations of this article
EID Journal Metrics on Scopus
Alex P. Salam1, Amanda Rojek1, Erhui Cai, Mihaja Raberahona, and Peter HorbyComments to Author 
Author affiliations: United Kingdom Public Health Rapid Support Team, London, UK (A.P. Salam); Centre for Tropical Medicine and Global Health, University of Oxford, Oxford, UK (A.P. Salam, E. Cai, M. Raberahona, P. Horby); Centre Hospitalier Befelatanana, Antananarivo, Madagascar (M. Raberahona); Centre for Integrated Critical Care, University of Melbourne, Melbourne, Victoria, Australia (M. Raberahona)

Cite This Article

Yersinia pestis, the causative agent of plague, is a Tier 1 select agent because of the high case-fatality rate associated with pneumonic plague and its potential as a bioterrorism agent in aerosolized form ( The death rate for persons with untreated primary pneumonic plague was reported to be almost 100% (1); the death rate for persons treated for primary pneumonic plague was 50% (1). Overall, the death rate for persons treated for primary pneumonic plague was high despite the sensitivity of Y. pestis to aminoglycosides, quinolones, and tetracyclines (2,3) and the relatively good penetration of some of these antimicrobial drugs into lungs (4,5). During the 2017 Madagascar pneumonic plague outbreak, the observed death rate for treated persons appeared to be substantially lower than that reported in the literature (6). Many articles that quoted a 50% death rate for treated primary pneumonic plague were cited in a 2000 study by Ratsitorahina et al. (7), which described a small outbreak in Madagascar in 1997. The article indicated that the data showed an overall death rate of 53% but did not state the number of deaths. However, the death rate for treated persons with confirmed or probable plague was 10%. On reviewing reports that cited Ratsitorahina et al., we identified 9 studies that referenced 50% of persons treated for pneumonic plague who died, 1 study that referenced 40%, and none referencing lower rates. One was a review cited 9 times about persons treated for primary pneumonic plague for whom the death rate was 50%. We identified 6 reports that stated but did not reference a 50% death rate for persons treated for pneumonic plague.

The Study

To address the lack of evidence supporting the frequently cited 50% death rate for persons treated for primary pneumonic plague, we conducted a systematic review and meta-analysis. We followed PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses, and MOOSE (Meta-analysis of Observational Studies in Epidemiology [8]) guidelines. The study was prospectively registered on PROSPERO (CRD42018086223) (

We searched PubMed and Embase covering 1946–2017 using the search terms “Yersinia pestis” or “plague” and “pneumon*” and limited our search to human data. We searched references and included articles describing death (within a 28-day period from illness onset) among patients with confirmed, probable, and suspected primary or undifferentiated (i.e., primary or secondary not distinguished pneumonic plague 1999 World Health Organization case definition, We did not restrict by study type, language, or minimum patient number.

Two authors reviewed and extracted data; a third author resolved any disagreements. Data fields extracted included year and country of the outbreak, number of patients who survived and died (stratified by antimicrobial drug status), number of patients receiving different antimicrobial drug classes, time to antimicrobial drug administration, and receipt of plague vaccination or prophylaxis (these patients were excluded). We calculated the risk from the number of events and participants in each group.

We performed a meta-analysis using a binomial-specific method. We assessed heterogeneity using the χ2 test and quantified results with the I2 statistic. In addition, we preplanned 2 sensitivity analyses to examine whether our estimation of death was influenced by the inclusion of specific articles (pneumonic plague was not confirmed as primary disease or patients with suspected and probable disease). We conducted statistical analysis using R version 3.6.0 (R Project,

We reviewed 362 articles (Appendix Figure 1). We described 1,107 patients in 44 articles (Appendix Table). Twenty-nine articles reported antimicrobial drug use in 108 patients with confirmed or probable pneumonic plague. For pneumonic plague patients receiving antimicrobial drug therapy, the pooled death rate was 17% (95% CI 8%–31%; I2 = 47%) (Appendix Figure 2). Pneumonic plague patients who did not receive antimicrobial drug therapy had a pooled death rate of 98% (95% CI 73%–100%; I2 = 47%) (Appendix Figure 3). Pneumonic plague patients for whom antimicrobial drug status was unknown had a pooled death rate of 46% (95% CI 32%–61%) (Appendix Figure 4). Heterogeneity was significant (I2 = 91%; p<0.01). The pooled death rates were similar when sensitivity analysis was conducted (Table). Antimicrobial drugs in the reports were aminoglycosides (90 courses), quinolones (24 courses), sulfonamides (22 courses), chloramphenicol (16 courses), tetracyclines (14 courses), and cotrimoxazole (3 courses). Six reports described time to from admission to antimicrobial drug administration, but the nonstandardized reporting precluded stratification by this measure.


Our meta-analysis identified a 17% death rate for persons treated for pneumonic plague, in contrast to the 50% often reported in the literature. The death rate for the 2017 Madagascar outbreak was published after we completed our systematic review but is consistent with our findings (25% in confirmed cases) (9). These figures compare with 13.6% for patients who died in the hospital of community-acquired pneumonia; 12.3% who died of Streptococcus pneumoniae infection; 14.7% who died of Legionella species infection; and 32%–61% who died of Staphylococcus aureus, Escherichia coli, Klebsiella species, or Pseudomonas aeruginosa infections (10). However, persons who died of other etiologic causes were predominantly elderly and had underlying conditions (10).

Our review indicated insufficient standardized data to stratify death by time from symptom onset to antimicrobial drug administration. The literature we assessed often stated that pneumonic plague is fatal in almost all patients who start antimicrobial drugs >24 hours after symptom onset. Generally, descriptions cite either 1 article, in which 11 patients treated within 24 hours survived and 2 treated after 24 hours died (11), or a handful of isolated case reports. However, case reports and series also exist in which patients survived despite starting antimicrobial drugs >24 hours after symptom onset (1214).

An accurate estimate of death is crucial for several reasons. First, it is helpful for public health planning during outbreaks, including the allocation of healthcare resources and the development of social mobilization campaigns. The commonly reported high death rate associated with primary pneumonic plague contributes to fear and panic among healthcare workers and the public. For example, anecdotal reports indicating concerns during the Madagascar outbreak were the following: healthcare workers taking continuous antimicrobial drug prophylaxis, mass public use of over-the-counter antimicrobial drugs, asymptomatic persons visiting the hospital, and sick persons avoiding the hospital. Accurate assessment of death is also essential for clinical trial design. For example, the required sample size would be 134 (power 0.8, α = 0.025) for a binary outcome superiority trial in which the death rate in the control arm was 50% and the intervention was assumed to reduce death by 50% (similar to the assumptions in a clinical trial of gentamicin vs. doxycycline in Tanzania in 2002) (15). However, a sample size of 476 would be required in a trial in which the death rate in the control arm was 20%. A sample size renders a superiority trial unfeasible. Even during the Madagascar outbreak, the largest outbreak of pneumonic plague this century, the final number of confirmed pneumonic plague patients was only 32 (9).

The major limitation of our meta-analysis is the sporadic reporting of clinical data and the relatively small number of cases for which antimicrobial drugs treatment status was described. Reporting bias in the literature also is likely, and pneumonic plague patients who survive are more likely than those who do not to be reported. Nonetheless, data indicate that the percentage of persons who die of treated pneumonic plague appears to be substantially lower than is frequently reported in the literature.

Dr. Salam is the clinical researcher for the United Kingdom Public Health Rapid Support Team. His primary research interests include clinical research in epidemic prone diseases. Dr. Rojek is senior clinical fellow at the Centre for Integrated Critical Care, University of Melbourne. Her primary research interests include clinical research in epidemic-prone diseases.



We thank Freya Shearer for assisting with statistical programming.

The UK Public Health Rapid Support Team is funded by UK aid from the Department of Health and Social Care and is jointly run by Public Health England and the London School of Hygiene & Tropical Medicine. The University of Oxford and King’s College London are academic partners. P.H. is supported by funding from the Department for International Development and the Welcome Trust (215091/Z/18/Z) and the Bill & Melinda Gates Foundation (OPP1209135).



  1. Prentice  MB, Rahalison  L. Plague. Lancet. 2007;369:1196207. DOIPubMedGoogle Scholar
  2. Centers for Disease Control and Prevention. Recommended antibiotic treatment for plague [cited 2015 Aug 25].
  3. Wendte  JM, Ponnusamy  D, Reiber  D, Blair  JL, Clinkenbeard  KD. In vitro efficacy of antibiotics commonly used to treat human plague against intracellular Yersinia pestis. Antimicrob Agents Chemother. 2011;55:37527. DOIPubMedGoogle Scholar
  4. Honeybourne  D. Antibiotic penetration into lung tissues. Thorax. 1994;49:1046. DOIPubMedGoogle Scholar
  5. Valcke  Y, Pauwels  R, Van der Straeten  M. Pharmacokinetics of antibiotics in the lungs. Eur Respir J. 1990;3:71522.PubMedGoogle Scholar
  6. Salam  AP, Raberahona  M, Andriantsalama  P, Read  L, Andrianarintsiferantsoa  F, Razafinambinintsoa  T, et al. Factors influencing atypical clinical case presentations during the 2017 Madagascar pneumonic plague outbreak: a prospective cohort study. Am J Trop Med Hyg. 2020;102:130915. DOIPubMedGoogle Scholar
  7. Ratsitorahina  M, Chanteau  S, Rahalison  L, Ratsifasoamanana  L, Boisier  P. Epidemiological and diagnostic aspects of the outbreak of pneumonic plague in Madagascar. Lancet. 2000;355:1113. DOIPubMedGoogle Scholar
  8. Stroup  DF, Berlin  JA, Morton  SC, Olkin  I, Williamson  GD, Rennie  D, et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA. 2000;283:200812. DOIPubMedGoogle Scholar
  9. Randremanana  R, Andrianaivoarimanana  V, Nikolay  B, Ramasindrazana  B, Paireau  J, Ten Bosch  QA, et al. Epidemiological characteristics of an urban plague epidemic in Madagascar, August-November, 2017: an outbreak report. Lancet Infect Dis. 2019;19:53745. DOIPubMedGoogle Scholar
  10. Fine  MJ, Smith  MA, Carson  CA, Mutha  SS, Sankey  SS, Weissfeld  LA, et al. Prognosis and outcomes of patients with community-acquired pneumonia. A meta-analysis. JAMA. 1996;275:13441. DOIPubMedGoogle Scholar
  11. McCrumb  FR Jr, Mercier  S, Robic  J, Bouillat  M, Smadel  JE, Woodward  TE, et al. Chloramphenicol and terramycin in the treatment of pneumonic plague. Am J Med. 1953;14:28493. DOIPubMedGoogle Scholar
  12. Begier  EM, Asiki  G, Anywaine  Z, Yockey  B, Schriefer  ME, Aleti  P, et al. Pneumonic plague cluster, Uganda, 2004. Emerg Infect Dis. 2006;12:4607. DOIPubMedGoogle Scholar
  13. Donaires  LF, Céspedes  M, Valencia  P, Salas  JC, Luna  ME, Castañeda  A, et al. [Primary pneumonic plague with nosocomial transmission in La Libertad, Peru 2010]. Rev Peru Med Exp Salud Publica. 2010;27:32636. DOIPubMedGoogle Scholar
  14. Luo  H, Dong  X, Li  F, Xie  X, Song  Z, Shao  Z, et al. A cluster of primary pneumonic plague transmitted in a truck cab in a new enzootic focus in China. Am J Trop Med Hyg. 2013;88:9238. DOIPubMedGoogle Scholar
  15. Mwengee  W, Butler  T, Mgema  S, Mhina  G, Almasi  Y, Bradley  C, et al. Treatment of plague with gentamicin or doxycycline in a randomized clinical trial in Tanzania. Clin Infect Dis. 2006;42:61421. DOIPubMedGoogle Scholar




Cite This Article

DOI: 10.3201/eid2610.191270

Original Publication Date: September 02, 2020

1These co–first authors contributed equally to this article.

The death rate for persons with treated pneumonic plague is often reported as 50%, but firm evidence for this figure is minimal. We conducted a meta-analysis of articles reporting the death rate for persons treated for pneumonic plague. The rate was 17%, substantially lower than the frequently cited 50%.

Table of Contents – Volume 26, Number 10—October 2020

EID Search Options
presentation_01 Advanced Article Search – Search articles by author and/or keyword.
presentation_01 Articles by Country Search – Search articles by the topic country.
presentation_01 Article Type Search – Search articles by article type and issue.



Please use the form below to submit correspondence to the authors or contact them at the following address:

Peter Horby, University of Oxford, Center for Tropical Medicine and Global Health, Roosevelt Dr., Oxford, OX3 7FZ, UK

Send To

10000 character(s) remaining.


Page created: August 05, 2020
Page updated: September 17, 2020
Page reviewed: September 17, 2020
The conclusions, findings, and opinions expressed by authors contributing to this journal do not necessarily reflect the official position of the U.S. Department of Health and Human Services, the Public Health Service, the Centers for Disease Control and Prevention, or the authors' affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.