ANTIMICROBIAL RESISTANCE

A POST ON ANTIMICROBIAL RESISTANCE

Global microbiology arrived in our small laboratory last week with the isolation of a Klebsiella pneumoniae, that appeared to be resistant to all antibiotics. It was from the sputum of a post operative patient, who had been in the hospital for two weeks.

Below is a photo of the disc sensitivity plate: 

What does this finding mean for a small rural hospital laboratory, and how did we deal with it ?

To begin with, it may be  useful to have an update on how increasing antibiotic resistance in Gram negative bacteria such as E.coli and Klebsiella spp has evolved over the last 10 years, and to understand some of the terms such as ESBL, Carbapenemase producers, metallobeta-lactamases,  VIM and NDM-1, terms you may have seen if you have seen in recent microbiology/infection journals.

Some readers of the post will know all of this, but there may be some who have not had the opportunity to keep up to date. So this discussion is mostly for the latter readers.

You will remember that beta lactamases are enzymes produced by many bacteria, that break down the structure of penicillin and similar antibiotics, so that the antibiotic can no longer kill the bacteria. Extended spectrum beta lactamases (ESBL's) are, as it says, beta lactamases with a spectrum of activity against a wider range of penicillin and cephalosporin antibiotics, so limiting the treatment options for infections caused by ESBL producing bacteria.

In the disc sensitivity testing, an E.coli or Klebsiella that has no zone to a third generation cephalosporin (cefotaxime, ceftazidime, cefixime etc), should be considered as an ESBL producer. A further test is required to determine if it definitely is an ESBL producer, one commonly used is the double disc test:

Method:

The double disc method compares the zone sizes of the isolate around a cefotaxime disc, and a combined cefotaxime plus clavulanic acid disc. A greater than 5mm diference between the zone size around the combined disc (the larger zone) and the cefotaxime disc confirms ESBL production.

The slide below shows a positive ESBL result.

While some rural hospital laboratories may be able to set this up, it may be better that the regional hospital has this facility, and smaller laboratories send suspected ESBL isolates to the regional laboratory.

 When considering antimicrobial resistance,  two components need to be considered:

a) the mechanism of resistance, eg a drug destroying enzyme such as a beta lactamase, the blocking of antibiotic uptake (tetracycline resistance), producing an alternative metabolic pathway (trimethoprim resistance)

b) the bacterial genes that code for the resistance mechanism, and how they are acquired by a bacteria.

One of the primary reasons why antibiotic resistance spreads both locally and regionally is that in many cases the genes coding for a resistance mechanism can be transferred between bacteria, and often a group of genes, coding for several resistance mechanisms, can be transferred together.

This is why we may see an E.coli that is an ESBL producer, and therefore resistant to penicillins and cephalosporins, and also resistant to aminoglycosides (gentamicin, amikacin), and to fluoroquinolones such as ciprofloxacin.

If such an E.coli was isolated from a blood culture, or a urinary tract infection, the options for treatment are very limited, and this is what is increasingly happening, even in rural tropical areas.  In many places, there would be no further antibiotic available, making treatment impossible.

The main group of antibiotics that can be used to treat such infections are the carbapenems, of which imipenem and meropenem are the main examples. Where these drugs are available, particularly in hospitals, they have been used to treat these highly resistant Gram negative infections.

However, a few years back, E.coli and Klebsiella strains were isolated that had also become resistant to carbapenems in addition to the other antibiotics. Investigations showed that they produced an enzyme, similar to beta lactamase called a carbapenemase, which inactivated the carbapenem antibiotic.

Below is a sensitivity plate with a Klebsiella pneumoniae resistant to imipenem and meropenem:

These new beta lactamases were found to have a metal ion in their molecule, and were called metallo-beta-lactamases. One of the first ones isolated and investigated was in New Delhi, India, and was labelled as New Delhi metallo-beta-lactamase-1, or NDM-1, a term you may have seen. Other metallo-beta-lactamases have been described, with names including VIM and OXA-48.

As with other resistance mechanisms, these are coded for by specific genes, which can move between bacteria. Therefore, an E.coli, which is an ESBL producer and resistant also to gentamicin and ciprofloxacin, could acquire the gene for NDM-1, and be resistant also to meropenem.

As with ESBL's, while the initial disc testing with resistance to meropenem/imipenem suggests a probable carbapenemase producer, further testing is required to confirm this. A special disc test (the Hodge test) can be used, but this is difficult to set up and monitor, and would more accurately be done at a regional laboratory. Specialised laboratories now use molecular tests (which we will talk about in post 11) to detect specific carbapenemase genes such as NDM-1.

Well, that's the end of the lecture ! 

So back to the question, if you start isolating resistant bacteria, whether ESBL producers or worse, what should the laboratory, and the hospital do about it?

Here is our experience.

The isolation of this multi-resistant Klebsiella made us realise that we didn't actually know how many times in the past we had isolated bacteria with resistance to most of our commonly available antibiotics, even if not as multi-resistant as this one. While we kept a day book of which bacteria had been isolated from which specimen (and date, patient details, etc) we had not been recording the sensitivity profile.

So, this was the first action. To begin a new day book (either an actual register, or on a spreadsheet on the computer), with columns in which to put sensitivity or resistance to the antibiotics tested. We in fact decided to have both a day book register, which the lab staff are used to using, but, as we had a computer with Excel on it, also to set up a spreadsheet. The spreadsheet  is updated from the day book at the end of each day. The advantage of the spreadsheet is that each month we can easily analyse how many isolates have been resistant to different antibiotics, how many are multi resistant. Also we can produce graphs and charts that we can use in a presentation to the doctors for them to be aware of the problem of antibiotic resistance. More details of setting up spreadsheets and databases to monitor infections will be given in post 9.

The second action was to be sure that our sensitivity testing was in fact accurate. We use the CSLI disc diffusion method. From a survey that was done in laboratories in district hospitals in our region last year, including sending round some test samples for comparison, it was evident that different laboratories were getting different results for the same isolate , and not all were following the correct criteria for the CLSI method that they said they were using.

One of the main problems was that rather than accurately measuring the zone diameter of "no growth" around a disc and comparing this to the standard table, laboratory staff were just comparing by eye zone sizes, writing down a "large" zone as "sensitive" and a small zone sometimes as "resistant" and sometimes as "partially sensitive".

This is illustrated in the image below:

We decided to look carefully at the CLSI protocol (ref link ) to ensure we were doing sensitivity testing correctly. You can check the link to assess your methods. The most important points to follow we realised are:

• Have some control organisms from the regional laboratory (E.coli, S.aureus, Pseudomonas aeruginosa), and do disc sensitivity with these and check against the CLSI control table that the zone sizes for different antibiotic discs with these control organisms are correct. This ensures that our stock of discs are potent, and our culture conditions are correct
• When testing clinical isolates, compare the zone sizes with the CLSI clinical samples table, and decide whether isolates are sensitive or resistant based on these sizes, not just by looking at the zone size.

A detailed description of the CLSI method is available at the following website:

http://www.microbelibrary.org/component/resource/laboratory-test/3189-kirby-bauer-disk-diffusion-susceptibility-test-protocol

An example of cut-off zone sizes for a range of antibiotics is given in the table below:

  This has been a rather long post! but it is an important topic, we have had to learn new things, and we hope it has been useful to you.

Further reading on Antimicrobial Resistance:

1. Increased multi-drug resistant E.coli from hospitals in Khartoum State, Sudan. Afr Health Sci 2012; 12:368-375.
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3557680/

2. WHO global report on antimicrobial resistance and surveillance.
http://www.who.int/drugresistance/documents/surveillancereport/en/

A LABORATORY ERROR

A CASE OF LABORATORY ERROR

The story:

An interesting and serious problem arose in the hospital over the past two months.

The father of one of the children in the children's ward  had had a persitent cough for the last two weeks. He visited on a daily basis. He was reluctant to go to the local health centre because he was busy harvesting his maize crop.  He was told he must have a sputum sample tested before he could visit again. Our senior laboratory technician was on leave, but the junior technician said he had been trained to perform ZN staining on a sputum sample.

A sputum sample was obtained, and processed in the laboratory. The technician saw the following down the microscope:

For answers to questions: See previous post: Answers to Lab Error Questions.

  

QUESTION 1:

This is a ZN stain of the sputum. What is seen on the slide?

The story continued:

The father had only visited his child in the ward, and had not been close to other children. His child, who was six years old, had had a BCG vaccination soon after birth. The father was referred to the TB clinic at the local hospital, and told he should not visit the unit until he had been on treatment for two weeks.

QUESTION 2:

From the above information, is it likely that any patients or staff in the ward were at risk of aquiring TB infection from the father? If so, give three points for action.

Five weeks later:

Five weeks after the incident, the laboratory cleaner did not report for work for a few days.  Her daughter brought a note saying that she had a fever and a bad cough. The previous night, the daughter said there was blood in the sputum. Four days later, her condition worsened and she was taken to the local hospital. The hospital laboratory examined s sputum sample and found it acid fast bacilli positive, and a diagnosis of pulmonary tuberculosis was made. Several staff remembered the cleaner, who also cleaned in the children's unit, had had a cough for about a week  before she took sick leave.

QUESTION 3:

TB is not uncommon among adults in this area, so it may be there is no link between the two cases. HOWEVER, do you have any suggestion how there may be a link between the two cases?

A third case occurs:

Two weeks later, the laboratory aid (who is responsible for cleaning the benches and discarding specimens) became unwell with fever and a cough. She attended the local health centre and a sputum sample wsas sent to the laboratory at the district hospital. It was found to be acid fast bacilli positive.

The children's ward sister is now concerned that somehow the ward or the laboratory was the source of infection in the cleaner and lab aid, and an investigation is begun.

Investigating an incident such as this has two main components:

a). To look for, and explain, epidemiological links between the cases

b). To investigate the processing of the specimen in the laboratory.

Had the lab aid or the cleaner come into contact with the first case?

There was no evidence that either the cleaner or the lab aid ever came into contact with the first case. It was therefore decided to investigate whether a breakdown in protocol had occurred in the laboratory, resulting in infection from the sputum sample.

These were the findings:

• the junior laboratory technician had performed ZN staining correctly and had worn a mask while handling the specimen
• the junior technician was busy with other tasks to do, and had left the specimen on the bench, rather than discarding it or putting in the refrigerator for storage
• when the lab aid cleaned the bench the following morning, she accidently knocked the specimen pot on the floor, which cracked and the lid fell off.
• the cleaner, who was in the laboratory at the same time, helped the lab aid to clean up the sputum, and rinsed the pot under a tap

QUESTION 4:

From the above findings, can you explain how and why the lab aid and the cleaner may have become infected?

The follow up:

Following the results of the investigation, a meeting was held with all the staff involved. The following actions were decided:

• The protocol for processing sputum specimens for ZN staining must be checked to ensure correct disposal of specimens is clearly described
• The junior technician to undergo a supervised programme with the senior laboratory technician for all laboratory activities and demonstrate his competence in the work
• Arrangements made for the temporary replacement of the lab aid and cleaner while they are on sick leave
• A  monitoring system set up to regularly audit the sputum processing in the laboratory
• The clinical staff are reminded to note any children in the t who develop a cough or fever to be referred to the paediatrician at the local hospital
• All staff told to report to a senior staff member if they develop a cough or unwell.

Both the lab aid and the cleaner made good progress on TB treatment, and fully recovered.

This episode was a useful experience: if even a small component of a laboratory protocol is not followed, serious infection hazards can result.

For the Medcare  programme, Uganda, the infection risk assessment proforma can be used to review the issues in this scenario.

The following link is to the form, with entries based on this case study:

Lab infection investigation form

Further reading on laboratory acquired infection:

1. TB risk among staff in a large hospital in Kenya. Int Jl TB and Lung Disease, 2008;12:949-954.

Free download at:http://www.ingentaconnect.com/content/iuatld/ijtld/2008/00000012/00000008/art00013?token=00511b8458b52781cca7b76504c48662a252c495b6c5f737b2d356a332b25757d5c4f6d4e227aaa5e

2. WHO Laboratory Bio-safety Manual.

Free download at: http://www.who.int/csr/resources/publications/biosafety/WHO_CDS_CSR_LYO_2004_11/en/

ANSWERS TO LAB ERROR QUESTIONS

ANSWERS TO THE QUESTIONS ON THE LABORATORY ERROR POST

QUESTION 1:

This is a ZN stain of the sputum. What is seen on the slide?

Pink bacilli (rod shaped bacteria), these are acid fast bacilli, typical of mycobacteria. While the sputum would have to be cultured to confirm the diagnosis, it is probably Mycobacterium tuberculosis.

QUESTION 2:

From the above information, is it likely that any patients or staff in the ward were at risk of acquiring TB infection from the father? If so, give three points for action.

The father is sputum positive, that means he is infectious with TB, and could infect anyone (patients or staff) who is susceptible, and who was close to him when he was coughing.

Action points could include the following:

§         Check if any children on the ward had not had BCG. Advice from the district TB officer should be sought. This would be particularly important if any child was known to be HIV positive, or had other reason for reduced immunity

§         All staff should be told to inform the senior hospital nurse if they develop a cough or become otherwise unwell

§          

QUESTION 3:

TB is not uncommon among adults in this area, so it may be there is no link between the two cases. HOWEVER, do you have any suggestion how there may be a link between the two cases?

The possible links (but we do not know the answers yet) could be:

a). The lab aid and the cleaner were both on the children's ward at various times when the father was there, and so became infected from him,

b). There may be no link, either with the father or the hospital. Both could have been infected in the community,

c). They may both have been exposed to infection in the laboratory.

QUESTION 4:

From the findings, can you explain how and why the lab aid and the cleaner may have become infected, but not the junior technician?

(These were the findings:

• the junior laboratory technician had performed ZN staining correctly and had worn a mask while handling the specimen
• the junior technician was busy with other tasks to do, and had left the specimen on the bench, rather than discarding it or putting in the refrigerator for storage
• when the lab aid cleaned the bench the following morning, she accidently knocked the specimen pot on the floor, which cracked and the lid fell off.
• the cleaner, who was in the laboratory at the same time, helped the lab aid to clean up the sputum, and rinsed the pot under a tap)

The junior technician wore a mask while processing the specimen, and apart from not discarding the specimen correctly, appears to have followed the correct protocol.

Both the lab aid and the cleaner were exposed to infectious material when the specimen pot fell and cracked, and in the process of cleaning up.

AN OUTBREAK OF MATERNAL SEPSIS

An outbreak of maternal sepsis.

One week  ago, JL, a 26yr old patient delivered (normal, vaginal delivery) a healthy female child in our hospital. She was discharged to her village the day after delivery.  Three days later a health worker came to the hospital reporting  that JL was very ill, with high fever and abdominal pain.  The baby was well.  We arranged for JL to be brought urgently back to the hospital. She was admitted into a single room in the maternity department.

QUESTIONS:

a. What do you think are possible diagnoses?

b.  What specimens would you take for the microbiology and haematology labs?

STAGE 1:

The patient had a fever of  40'C,  a low blood pressure, severe abdominal pain,  and an odourless vaginal discharge.

While the patient was being admitted, two things happened rather quickly: a health worker from another village came in a hurry to the hospital and said a woman BS who had delivered in the hospital five days ago was very ill, with symptoms sounding similar to those of JL. Also, a nurse in the maternity ward reported that a mother AK who had delivered two days ago, had spiked a temperature of 39'C,  and had abdominal pain.

QUESTIONS:

c. What do you think is going on?

d. Should someone assess the baby of JL and the other patient in the village?

STAGE 2:

The following specimens were taken on patient JL:

Blood sample for Hb and blood count

A urine specimen for culture and sensitivity

Blood sample for blood culture and sensitivity

Vaginal discharge sample for culture and sensitivity

Vaginal discharge samples were also taken from patients BS and AK.

LABORATORY RESULTS:

1. Hb 9.6 g/dl. White cell count 22,000, 90% polymorphs.

2. The findings in the microbiology lab are shown below:

Microscopy of the vaginal discharge:                     Culture appearance of vaginal discharge               

(Gram stain)                                                                         on blood agar

                                                     

 

QUESTIONS:

e. Describe the Gram stain and the culture appearances

f. What is the possible bacteria causing the discharge, and the infection?

THE STORY CONTINUES:

There are three patients with the same infection, all apparently originating from the maternity unit. This suggests a small outbreak, and cross infection.  As well as diagnosing the individual patients, we need to investigate the outbreak, and hopefully stop further infections.

INVESTIGATING AND CONTROLLING THE OUTBREAK :

(Remember Semmelweis, 1847. ..Google this if you are not sure).

While post partum sepsis may arise from bacteria the mother was carrying, as there are three linked cases, it suggests infection is being transmitted between patients, or from a common source; how might this happen?

We need to look at the practices and other factors associated with the maternity ward.  These are some of the questions to ask:

1. Is common equipment used between patients, eg a speculum?, how is it cleaned between patients?

2. Were the three cases in the ward at the same time.

3. Do delivery nurses/midwives wear gloves for examinations and deliveries? are these gloves changed between patients?

4. Are the hand washing facilities on the ward adequate? are water and soap always available?

5. Is the general hygiene of the ward (floors, bed sheets, patient washing etc) adequate.

FINDINGS:

 a. All three mothers had been in the same ward for at least one day together, JL and BS post delivery, and AK the day before delivery

b. Nurses were certain that they changed gloves between patients, but observation showed that when the ward was very busy and several mothers ready to deliver at the same time, this did not always happen

c. There were only three specula on the unit, but they seemed to be cleaned appropriately between patient use

d. The general hygiene on the ward seemed to be adequate although resources were limited.

e. Although there were sinks and soap , hand washing practices could have been improved:

  http://www.who.int/gpsc/tools/Five_moments/en

INTERPRETING THE FINDINGS:

The cause of the infections was Streptococcus pyogenes, Group A Strep.

This is a Gram positive coccus, appearing in chains, and producing beta (clear) haemolysis on culture on blood agar.

Fortunately Gp A streps remain sensitive to penicillin.

It is probable that one patient was colonised vaginally with Gp A strep, and this was transmitted to the other two patients by the unchanged gloves of one of the staff.

LESSONS:

a. Maternal (puerperal) sepsis is not uncommon in low resource areas, is potentially fatal, but can respond to penicillin if treated early

b. While the lab diagnosis was not essential for managing the individual  patients (though it did confirm the diagnosis), it was essential to show that cross infection had occurred.

c. In district hospitals in 2015, wherever they are located, microbiology labs should be able to do culture and sensitivity

d. If there appears to be a cluster of infections, or an outbreak, some basic epidemiology related questions must be asked: who is affected, where were they, when did they become ill, what factors may have caused transmission, are basic infection control measures adequate?

e. Based on the above, changes must be implemented to improve infection prevention and control

Link to Tropical Microbiology Laboratory website for more details on lab methods:

https://sites.google.com/site/tropmicrolaboratorynetwork/

Further reading on maternal sepsis:

1. Maternal and early onset neonatal bacterial sepsis: burden and

strategies for prevention in sub-Saharan Africa.

Lancet infectious diseases 2009; 9: 428-438.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2856817/#__abstractid2483334title

2. Microbial profile in women with puerperal sepsis in Gadarif State, Eastern Sudan

   Annals of Tropical Medicine and Public Health, 2013;6: 460-464.

http://www.atmph.org/article.asp?issn=1755-6783;year=2013;volume=6;issue=4;spage=460;epage=464;aulast=Ahmed