Print Friendly Version  


Preserve Antibiotics for the Future
How Epidemics of Untreatable Infections Develop
Reduce Antibiotic Use to Delay Antibiotic Resistance

Written by Thomas F. O'Brien, MD, Vice President, APUA; Associate Professor of Medicine, Harvard Medical School

The articles were made possible with the support of The Pew Charitable Trusts.


Bacteria have been evolving for three billion years into countless kinds totaling a billion trillion trillion with more mass than all animals. A few kinds, pathogens, can invade and infect tissues of animals, but few of them are infecting at any time. Yet those infecting bacteria have shortened human life more than all other causes.

Discovery of antibiotics that diffuse through human tissues, kill bacteria infecting them and cure the infections was the greatest medical advance ever. After each new antibiotic was widely used, however, bacteria increasingly became resistant to it and a new antibiotic was needed to cure the infections they caused.

Killing infecting but fewer non-infecting bacteria by treating truly infected patients while needed has been hard to attain case-by-case in a complex world. Ending antibiotic use for animal growth promotion is a cost-effective way to slow the progression of resistance before it outruns drug discovery.


Trouble was identified soon after the discovery of penicillin in 1942. Some early penicillin workers noted strains of bacteria that were not killed by it, and within a few years such strains were shown to produce an enzyme that actually destroyed penicillin. The significance of this may have been overshadowed over the next decade because a large number of new classes of antibiotics; aminoglycosides, tetracyclines, chloramphenicol and macrolides, were discovered and brought to market, and it seemed like there would be no end to them.[2]

But early in the following decade an end became apparent, and it was of two kinds. No further classes of antibiotics had come to market, and in fact only a few more would in the following half century. Worse was the realization that a particular strain, one of the most dangerous pathogens, Staphylococcus aureus, was spreading through the world infecting and killing people regardless of which of the new antibiotics were used to treat its infections. It had become resistant to all of them and rampaged for a decade until semi-synthetic penicillins and cephalosporins came to market in 1960.

Strains resistant to those, methicillin-resistant Staphylococcus aureus (MRSA), also emerged soon in some parts of the world but not in others, including the US, for another 15 years. Treatment for those now ubiquitous MRSA has depended on the difficult-to-use vancomycin and two recent drugs, with resistance to each now being reported. At the same time, multiple species of a different category of pathogen, Gram-negative bacilli, were also becoming resistant to more antibiotics and many to all of them becoming untreatable. The semi synthetic penicillins and cephalosporins were of limited help for this problem and so it continued to grow until the introduction of gentamicin in 1970. Gentamicin was effective for virtually all such strains, but only a short reprieve. In 1975, enzymes inactivated the newly introduced gentamicin. It became apparent that this, as well as earlier types of resistance, was spreading in global epidemics, not just of resistant strains of bacteria but of the smaller and more portable plasmids. These latter mobile genetic elements could carry genes expressing resistance from one strain of bacteria to a different strain, or even one of a different species, so that this new and previous susceptible strain and all its progeny were, thereafter, resistant [1].

Finally, around 1980 there was a triple rescue from the growing menace of multiple-resistant Gram negative bacilli, from modifications of older classes, giving us newer generations of cephalosporins, fluoroquinolones and carbapenems. Unfortunately, due to natural bacterial mutations and to human misuse, resistance to the first two has been building steadily and merging, and in recent years to the third also, so that untreatable strains are again spreading and killing with no definite rescuing agent in sight [2].

As antibiotic resistance increases, it is likely to be more devastating, especially for the larger part of the world that lacks access to laboratories to identify what antibiotic would still cure which infection. These infections will have to be treated blindly and more often ineffectively from a limited stock of locally available antibiotics.

In retrospect now, it appears that when each new antibiotic was developed and marketed, the resistance genes that would limit or end its usefulness already existed. Its useful life was predestined by the time it took for one or more of those genes to either arise in pathogens in multiple places by local antibiotic usage selecting mutant resistant genes, or for existing resistance genes to be mobilized eventually from obscure bacteria somewhere and then spread widely from that place or those places.

Antibiotics were introduced at varying intervals and resistance to each emerged at irregular intervals thereafter. Sometimes a new rescuing antibiotic arrived before the wide spread of infecting strains[3] that were resistant to all earlier antibiotics - and thus untreatable even in the affluent world - or of strains resistant just to antibiotics commonly available in the rest of the world - and thus virtually untreatable by the blind therapy there [3]. And sometimes a rescuing antibiotic for either part of the world did not arrive in time.

The antibiotic era can be seen now as a double miracle. It seemed like a miracle when it began that there could be such things as antibiotics – and, indeed, they were commonly referred to then as “miracle drugs”. And we might now consider it wondrously fortunate that it took the immensely resourceful microbial world a decade or more to come up with and distribute ways to nullify each one. Stringing together a series of such intervals of delayed response, with occasional lethal gaps, was how the antibiotic era has survived. Having an effective antibiotic came to seem an entitlement, but was really just luck –and luck is running out.



The resistance genes that made infecting bacteria resistant to a new antibiotic after it was used had existed before. Some were mutants of infecting bacteria, unseen until the antibiotic selected for their overgrowth. More took years to first get from obscure to infecting bacteria somewhere and to then spread from there.

Antibiotics drive resistance by killing susceptible bacteria and letting a resistant one then multiply explosively to replace a billion dead susceptible ones overnight. After three billion years of evolution, few infecting bacteria had any resistant gene, but 70 years of antibiotic use has put many into more than a quarter of them.

A succession of irregular global epidemics has given us a growing global catalog of resistance genes and growing local inventories of them in any region that might diminish, but could be rapidly recalled. The examples cited herein glimpse the enormous resources of the world’s bacterial populations for responding to introduced antibiotics and the negligible barriers to those responses of the species of the bacteria or the species of the hosts that carry them. This global web necessitates the need for a global surveillance system so that the medical community is never caught off-guard by emerging resistances. Most importantly, it is imperative that use of antibiotics in agriculture be restricted to therapeutic use only, and that antibiotic use for growth promotion and other non-therapeutic uses be terminated.

Antibiotics are small molecules that enter bacterial cells and bind to and block specific functional sites within them. Some antibiotic resistance mechanisms are alterations in a functional site that make antibiotic binding impossible. Resistance can also emerge through bacterial enzymes or other proteins that destroy the antibiotic, nullify its effect, or eject it from the bacterial cell. Either type is expressed by what can be called a resistance gene, which susceptible cells do not have.

Naomi Datta found almost no resistance genes in infecting bacteria stored from the 1930s, and resistance to each new antibiotic was rarely seen, until after it had been widely used for years or even decades [1]. So the central questions are, how do resistance genes arise and spread and how can we delay it.

Some resistance genes emerge from mutation
Resistance genes can now be seen to emerge in two different ways. Resistance to some antibiotics, such as fluoroquinolones, may arise in a strain of bacteria from a mutation in a gene that is normally carried by strains of that species. This classic Darwinian model has been in our thinking about antibiotic resistance for much of its seventy years. Such mutations have been occurring in those strains at rates like once in a million cell divisions, but they are entirely unnoticed until the antibiotic is used. In a crowded bacterial world, a single resistant bacterial cell, with no other advantage, is condemned to obscurity, until the antibiotic arrives to kill its billion susceptible neighbors and give it some space - and let it then overgrow by such “selection”.

More resistance genes emerge unpredictably from obscure recombinant events
It is increasingly recognized now that many of the most important resistance genes, including those for MRSA, vancomycin-resistant enterococci (VRE), penicillin-resistant Streptococus pneumoniae (PRSP), etc. differ too much from genes in their susceptible ancestors to have arisen by mutation from them. Such resistance genes, moreover, can often be seen to have first emerged only after many years of use of the antibiotic (30 years for PRSP or VRE) and initially, in one or a few parts of the world - with gradual but eventual spread to other parts.

Finally, some of these same genes are now being found in more obscure kinds of bacteria. Many, if not most, of the most damaging antibiotic resistance genes can thus be seen to have emerged not from simple, predictable mutation in the strains and species in which they are now widely infecting. The resistance genes appear, instead, to have emerged by one or a succession of rare mobilization and recombination events from some remote strain or species, which had the gene. The strain/species was obscure, because it was outside of the small subset of pathogens that clinical laboratories report. It may take years for such a resistance gene to first emerge somewhere in infecting bacteria that those laboratories would notice, and then more years for it to spread to other places.

Examples of the evolution and spread individual antibiotic resistance genes
An enzyme that destroys penicillin and ampicillin, TEM, spread widely among Gram-negative bacilli in humans and animals throughout the world. This enzyme accounted for most of the resistance to ampicillin in the commonest infecting Gram-negative bacillus, Escherichia coli, when 15% of them were resistant to ampicillin 35 years ago. Now, more than twice that many are resistant and, consistent with what Naomi Datta found, that extrapolates back to zero when penicillin use began.

At about the midpoint of the era, penicillin could still treat all gonorrhea, and ampicillin could treat any kind of bacterial meningitis anywhere in the world. So both of these major diseases could be treated optimally immediately without a laboratory. This was a huge advantage anywhere but immense to the larger world, that has more of those diseases and fewer, or no, laboratories. Within a few years, however, E. coli carrying resistant enzymes had presumably become prevalent enough and contacted enough of these very different kinds of bacteria to engineer a genetic transfer of this resistance to a gonococcus and another to the commonest cause of meningitis, Hemophilus influenzae. Such “second generation” resistant strains then spread to cause untold numbers of treatment failures and deaths.

When entirely new cephalosporins were introduced around 1980, all Gram-negative enteric bacilli were susceptible. Within 5 years, however, mutations enabled TEM and the others to “open wider” and destroy these drugs too. Four decades of penicillin and ampicillin use had thus spread the resistance enzyme gene from somewhere to make resistance to both prevalent everywhere, but also to be available widely for “third generation” mutations to destroy the new drugs that were meant to fix the problem.

Vancomycin, an entirely different antibiotic for entirely different kinds of bacteria, Gram-positive cocci, was also used for many decades. Unlike penicillin or ampicillin., however, no one saw growing resistance to it during that time or any resistance at all. Then about 20 years ago, such resistance appeared first in enterococci of animals in Europe that had been given an analog of vancomycin and then by the same mechanism in patients in hospitals all over the United States [2].

This mechanism, too, has had a beginning “second generation” in 8 patients in whom it transferred vancomycin resistance to MRSA, although none have yet spread to additional patients - a truly catastrophic threat [3].

These “second” and “third generation” resistance examples tell us that resistance begets more resistance and that there is no going back. A succession of such irregular global epidemics has given us a growing global catalog of resistance genes and growing local inventories of them in any region that might diminish but could be rapidly recalled [4]. The examples also glimpse the enormous resources of the world’s bacterial populations for responding to introduced antibiotics and, importantly, the negligible barriers to those responses of the species of the bacteria or the species of the hosts that carry them. This global web necessitates the need for a global surveillance system so that the medical community is never caught off-guard by emerging resistances.


Antibiotic use drives each step in the emergence, spread and evolution of resistance genes. It selects and spreads mutants. It amplifies obscure bacteria with resistance genes and genetic events that bring those genes to infecting bacteria. It amplifies resistant bacteria in treated hosts to be the onesthat then infect them or spread to the next host in the chain of spread.

The progression of antibiotic resistance seen as growing a variety of resistance genes, prevalence of resistant bacteria and prolonged sickness or death after treatment failure, seems increasingly a function of total number of bacteria killed by antibiotics, with little distinction for the species of the bacteria or their hosts. It is imperative that human beings slow the progression of resistance to increase the chances that rescuing antibiotics will arrive in time. Eliminating use of antibiotics for growth promotion and other non-therapeutic purposes (feed efficiency, poor on-farm hygiene) in food animal production would be a significant step toward preserving this precious resource.

Oil that could be channeled into valuable specific uses has now been gushing into the Gulf of Mexico for over two months. It is quietly wreaking environmental havoc, and no one has yet been able to stop it. Antibiotics, similarly, have an extremely valuable specific use. By diffusing through sterile body tissues and killing any bacteria invading any of those tissues, they have extended human life more than anything else has. Outside of such specific use in infected tissues, however, they wreak a growing and dangerous environmental havoc, gushing into the general community for seventy years.

Bacteria have been evolving and diversifying for three billion years into every niche on the planet. They are excluded mostly only from larger organisms like humans and other animals and then only temporarily, now totaling a million trillion trillion (1030) bacteria of thousands of different kinds living in finely balanced ecosystems all over the world. Accordingly, the discovery 70 years ago of antibiotics able to diffuse through living tissues, kill any bacteria infecting them and save all those lives can be seen as the most valuable medical advance of all time.

The crowded lives of bacteria
Bacteria are very tiny, and one needs a 1000X magnification to see them. However, they make it up by their numbers to a mass perhaps greater than that of all the big animals combined. They have gained their huge populations not only by penetrating almost everywhere but also by packing tightly in intricately synergistic communities. A cubic millimeter of stool may contain a trillion bacteria of hundreds of kinds, and fertile soils have only a few logs less. Adjustments in such populations are made dynamic by the explosive multiplication rates that bacteria can attain, with generation times as short as twenty minutes.

The “miracle” of manufactured antibiotics
Seventy years ago, humans discovered and began to manufacture antibiotics, eventually tons and tons of many different kinds. The intent was to have small amounts of an antibiotic diffuse through the sterile tissues of infected patients, kill the bacteria that were invading, and cure the infections and save the patient’s lives. Everyone was so astonished that such a thing was possible, that for a while it was the only aspect of antibiotic use that one could think about. Where else it was going, one could care less.

Antibiotics were portable, moreover, and soon cheap and ubiquitous. Earlier control of infection had been by prevention, which meant approaching everyone, while they were healthy and not feeling any immediate need for it, with big programs of total immunization or clean water that needed advance planning and large appropriations. No one needed an antibiotic until and only when his or her child was infected. And then they were plenty motivated to get one, so distribution could be marketplace, rather than public health. Few new products had ever achieved worldwide availability so quickly.

The extraordinary power of an antibiotic to amplify and spread resistance to itself derives from its lethality for enormous numbers of susceptible bacteria, and the speed with which a bacterial cell resistant to it can multiply to replace all the susceptible bacteria. The power of an antibiotic to amplify and spread resistance to other antibiotics is due to the accumulation in bacteria that have survived them and of the transferable mobile genetic elements in those surviving bacteria [1].

Antibiotic use drives each step in the emergence, spread and evolution of a resistance gene. It uncovers, amplifies and spreads resistant mutants, otherwise doomed to solitary obscurity. It amplifies obscure strains with prototype resistance genes, making them more available to recombinant events that may take a decade or more to bring them to infecting bacteria. It amplifies the numbers of a resistant strain colonizing any human or animal carrier that carries one, and so increases the chances that it will be the one that eventually infects that carrier or gets transferred to the next carrier in the chain of transmission.

After three billion years of bacterial evolution, the world’s infecting bacteria had almost no antibiotic resistance genes, but a half century of antibiotic use then spread many into more than a quarter of them [2]. People or animals receiving antibiotics as well as countries that use more antibiotics have been found to have more antibiotic resistant bacteria [3]. From all we know, the progression of antibiotic resistance would appear to be ultimately some cumulative function of how many bacteria have encountered an antibiotic.

If all of this had been known 70 years ago, better control of antibiotics might have been established to slow the amount gushing harmfully onto non-infecting populations of bacteria. It is difficult, however. Antibiotics that diffuse through infected tissues are also excreted into skin and gut where they encounter thousands of times more non-infecting bacteria, and in animals then also into a bacteria-rich environment. Uncertain diagnosis and excess caution combine to treat with antibiotics more people or animals who have no infecting bacteria than who do, and to treat those who do for too long. All of these increase the number of bacteria exposed to antibiotics - the basic metric for this gushing. Programs to minimize each of these types of unnecessary use have succeeded, but are hard to establish widely.

In this context, giving antibiotics to food-production animals solely to promote their growth seems their least essential use and its cessation our best opportunity to slow the gushing. Since it is a large application involving many non-infecting bacteria its ratio of non-infecting to infecting bacteria is presumably infinite. The claimed benefit is a small reduction in meat cost, which would not appear to weigh heavily against the hazard of further antibiotic resistance, particularly in a country campaigning against obesity.

Looking back over 70 years at parallel timelines for the introduction of antibiotics and for the emergence and spread of resistance to them, one has seen both in-time and not-in-time arrival of rescuing antibiotics. One needs to slow the progression of resistance to increase the chances that rescuing antibiotics, with all the uncertainties of their development, will arrive in time. The growing evidence sketched here indicates that this requires an overall reduction in the numbers of bacteria that are being exposed to antibiotics.

Only by employing a multi-pronged approach to this serious public health problem can one hope to make an impact on preserving this precious resource to both safeguard and extend human and animal life. Part of the solution lies in reducing antibiotic use in circumstances that do not require them. Inappropriate/over use of antibiotics in food animal production is a case in point. It is essential that use of antibiotics in agriculture be limited to the treatment of diseased animals and should not be used for non therapeutic purposes: growth promotion, feed efficiency, or to compensate for stress of transport and on-farm conditions of crowding and poor hygiene [4, 5]. Use of alternative infection prevention measures is encouraged, where possible. Fluoroquinolones and third generation cephalosporins, antibiotics critical to treating human diseases, should be restricted to treating refractory infections in individual animals [4]. In addition, antibiotics should be administered to animals only on prescription by a veterinarian [4].

To assess the human health risk and inform public health policy, quantitative data on antimicrobial use in agriculture should be made available by pharmaceutical manufacturers, importers and end users [4]. Regulatory agencies should consider the ecology of antimicrobial resistance –the processes of spread and complex interactions between bacteria – both pathogens (disease causing) and non-pathogens (commensals), food animals, humans, and their environments [4].  Surveillance programs for antimicrobial resistance should be harmonized to permit integrated analysis of human and animal data [4].

1. Watanabe, T., Infective heredity of multiple drug resistance in bacteria. Bacteriol Rev, 1963. 27: p. 87-115.
2. Pournaras, S., et al., Clonal spread of KPC-2 carbapenemase-producing Klebsiella pneumoniae strains in Greece. J Antimicrob Chemother, 2009. 64(2): p. 348-52.
3. Archibald, L.K. and L.B. Reller, Clinical microbiology in developing countries. Emerg Infect Dis, 2001. 7(2): p. 302-5.
4. FAAIR, Policy recommendations. Clin Infect Dis, 2002. 34 Suppl 3: p. S76-7.
5. Institute of Medicine. in Microbial Threats to Health: Emergence, Detection, and Response. 2003: National Academy Press.  

These articles were published in the APUA 2010 Newsletter, Vol. 28 No. 2

2014 Alliance for the Prudent Use of Antibiotics | 136 Harrison Ave, M&V Suite 811, Boston, MA, 02111 | Tel 617.636.0966 | apua@tufts.eduHome | Search | Contact Us