Emerging Infections of International Public Health Importance
Workup on the Unknown—Lab Investigations
Romesh Gautom, PhD
Dr. Romesh Gautom is the Director of the Washington State Public Health Laboratories (PHL). At the PHL he has established a strong track record for developing innovative new methods to investigate diseases of public health interest. Of particular value has been a new rapid DNA fingerprinting procedure based on Pulsed-Field Gel Electrophoresis, which has formed the backbone of the national PulseNet system.
A brief overview of Washington State Public Health Laboratories
The role of the public health laboratory (PHL) is disease identification and outbreak investigation, reference services, specialized testing, and environmental testing. In addition, we conduct rapid testing, laboratory improvement, applied research, support of surveillance and epidemiology investigations, and emergency preparedness and response. The PHL works with a number of partners: Communicable Disease Epidemiology, local health departments, STD program, TB program, Food program, Radiation program, other laboratories and hospitals, federal agencies (CDC, FDA, etc.), and state colleges and universities. There are three major sections: newborn screening, environmental laboratory sciences, and public health microbiology. There are two support sections--operations & training and safety & quality assurance--that provide technical support to these three sections.
In our newborn screening program, every single child born in Washington is tested for: congenital hypothyroidism, phenylketonuria (PKU), congenital adrenal hyperplasia, sickle cell disease, galactosemia, and biotinidase deficiency. We have 75,000 to 80,000 births in the state annually, and all babies have their heel pricked and the blood is tested in our laboratory. Thanks to this newborn screening program, this past year, 60 babies were spared from mental retardation, physical disability, and death.
The second section is environmental laboratory sciences. Here, we perform tests for drinking water quality, reference work for the state, radiation, and outbreak investigation work. This section is composed of environmental microbiology, environmental chemistry, and environmental radiation chemistry:
For our operations & training program, we have become heavily involved since the 9/11 terrorism attack. We work with a variety of laboratories under local/county health departments, hospitals, and physician offices. We provide training for gram stain, urinalysis, parasitology, blood cell morphology, and bioterrorism preparedness and response.
For our safety and quality assurance program, we develop and implement a geographically unique quality assurance, safety and health program for the public health laboratories in Washington State. The program assures that our safety officer works in a safe environment and is performing good quality control and assurance.
Finally, in public health microbiology, we perform routine testing; reference work, disease surveillance and outbreak investigations, and the development of new methodologies. Routine testing is done by collecting samples obtained from the physician office and sending them to the public health laboratory for work up. For example, we test Mycobacterium tuberculosis (the TB laboratory serves as the state reference laboratory), STD’s (syphilis, HIV, and Chlamydia) and conduct routine testing of rabies in animals. Hospitals have their own laboratories, but suspected bioterrorism samples and samples for antibiotic susceptibility testing come to our laboratory. Samples from county clinics, prisons, and family planning clinics also come to our laboratory.
Reference work and disease surveillance go hand in hand. In our enteric laboratory we deal with enteric pathogens such as E. coli, salmonella, shigella, and campylobacter. Submitters of samples include local health departments, hospitals, and private laboratories. Specimen types include clinical (e.g. stool), environmental (e.g. duck feces), and cultures from other laboratories for confirmation (e.g. samples from hospital laboratories). We perform biochemical and serotyping tests; turn around time (TAT) can range from 3 days to 3 weeks. We perform 2500 samples/year, and some samples are sent to CDC for additional workup. When the large E. coli outbreak occurred in our state in 1993, the samples were all processed here.
The reference laboratory confirms and identifies cultures isolated by other laboratories. Submitters include local health departments, hospitals/physician laboratories, and private laboratories. This includes testing for unusual pathogens, such as anthrax, plague, tularemia, and botulism. We perform a variety of biochemical and serologic tests, such as fatty acid analysis, antimicrobial susceptibility tests, legionella tests, and bioterrorism agent identification. The workload ranges from 50 - 60 samples/month, and the TAT depends on the organism.
The Public Health Laboratory and outbreak investigations
An outbreak or an epidemic is the occurrence of more cases of disease than expected in a given area or among a specific group of people over a particular period of time. When we have an outbreak, the three sections of epidemiology, environmental health and the laboratory collaborate to find the source.
Epidemiology would consider the incubation period (if known), the duration of the illness, the severity of the illness, and the symptoms (and how long the symptoms lasted).
This slide shows you the number of food-borne outbreaks in Washington State in 2002.
Campylobacter is very frequently isolated in our state, and it is associated mostly with chicken.
Outbreaks can spread if not controlled. To prevent future outbreaks, we have to identify the organism (s) and source (s). After we have verified the agent and the source, we need to find the mode of spread and track the outbreak. This is a collaborative effort between environmental health, epidemiology and the public health laboratory.
There are some traditional methods and some advanced molecular methodologies that we use when handling outbreaks.
The traditional methods are often still gold standards, but they generally take more time. That is why more public health laboratories are shifting toward molecular methodologies in conjunction with traditional methodologies.
Molecular diagnostics and molecular epidemiology
The basic principle of molecular diagnostics is the detection of a specific nucleic acid sequence by hybridization to a complementary sequence or probe followed by detection of the hybrid. There are several nucleic acid amplification based methods.
The main methodology is polymerase chain reaction (PCR). A newer method of PCR, real-time PCR, is becoming popular. New methodologies that are under development in the PHL can enhance our capabilities to respond to public health needs.
Let’s assume that you suspect there is E. coli 157 or bordetella pertussis in your sample. You develop primers based on either the gene you are going to target or the portion of that gene. The primers are very specific for that gene or the portion of that gene, so you take a primer (s) and mix them with the DNA from the sample. Then you add DNA polymerase, DNTP and nucleotides in the mix and amplify the target gene. You can set the machine to run for a certain number of cycles. After one cycle, one gene would be amplified to two. After two cycles, to four, and after 20 cycles, it becomes over 20 million copies
If the gene for the target was present in your sample, you would see it on a gel.
With real time PCR, we have a system called TaqMan, where you do not have to run the sample on a gel—it is less labor intensive and time consuming. The machine uses two primers; you have a probe that is attached with two dyes, a photo dye and a quencher dye. Your probe attaches on your target, gets amplified, and after about two hours, this is what you see on the TaqMan.
The computer can determine the presence of the target gene.
Specific procedures are followed in order to identify an unknown gastrointestinal pathogen.
The sample is sent to our bacteriology laboratory, where they culture and look for salmonella, shigella, E. coli and other bacterial pathogens. This is a fairly rapid method, and biochemical tests can be used for confirmatory tests. You can also send a portion of the sample to our virology laboratory, where they perform PCR or other tests. We can do nucleic acid amplification in four hours using TaqMan and identify the viruses. Some pathogens require special requirements--campylobacter, for example, is an anaerobe and we have to use special media to identify it.
When we have some unusual pathogens in our sample, we try to revive the pathogen by incubating it in blood culture.
We can do 30 different types of biochemical tests, and if it is still indeterminate, we do 96 biochemical tests in an automated reader called the Biolog. The 96 wells all have different substrates and after adding the sample, you look for color change for identification. If the organism is still indeterminate, then Midi (gas chromatography) is used, where you look at the lipid profiles of the pathogen’s cell wall. Sometimes we use pulsed field gel electrophoresis (PFGE) and DNA sequence analysis. If still indeterminate, we send the sample to CDC for further tests.
Recently there has been a heightened awareness of SARS. If we have an unknown respiratory pathogen in a suspected SARS victim, we first use RT-PCR TaqMan.
The sample can be tested for SARS within 6 hours, and once it’s negative, we do testing for influenza A, influenza B, hMPV, adenovirus, RSV, and some bacterial pathogens. If negative, we also do further serological testing for confirmation (EIA-linked antibody testing to IgA, IgG and IgM).
SARS laboratory testing procedures and other methods are developed by the CDC, and standardized procedures are disseminated to all PHL’s. They provide comprehensive information for quality control and quality assurance, and provide technical support and training to ensure comparability of tests. We also have annual meetings where we share each others’ existing and newly developed methodologies. In addition, through the Association of Public Health Laboratories, all the state laboratory directors get together at least once a year to share information and to ensure standard protocols.
We do EIA for IgM and IgG ELISA testing on the human serum for West Nile virus and St. Louis Encephalitis virus.
If the results are positive the sample is sent to CDC for plaque reduction neutralization test (PRNT), and they can confirm if the sample is West Nile, St. Louis encephalitis, or some other flavivirus.
Bordetella pertussis has been a major problem in King County and in other counties in Washington. The gold standard and traditional way of detecting bordetella pertussis is by culture, which can take up to 6 to 8 days. DFA can be done faster, but it is not very precise and leads to many false positives. We wanted to come up with a more robust, specific, and rapid test for bordetella pertussis, so we developed a PCR test.
We have a primer set to look for human DNA to make sure that the sample that you collected was a good sample. There can be a problem if the sample did not come from a good swab and the results come out negative. However, your sample could have been colleted improperly. So whenever we test for bordetella pertussis, we also look for the human DNA band to make sure that the swab picked up epithelial cells from the nasophayrngeal cavity. If this is negative and the human DNA is not amplified by PCR, the result is indeterminate and we need to recollect the sample.
Norovirus is the most common cause of nonbacterial gastroenteritis. Uncooked/undercooked shellfish, salads, cold foods, bakery products, ice, swimming water, etc. are the vehicles for transmission. They cannot be grown in any culture media, and serology tests are still in the developmental stages. The gold standard so far is EM (electron microscopy), but it is tedious, cumbersome, and expensive. PCR, however, provides us with sensitive, specific, and rapid diagnosis. We developed an RT-PCR method, where we extract the RNA, convert it to DNA, and amplify the DNA for detection of the band of interest.
We have all the primer sets for the many different genotypes of noroviruses.
Similarly we have a PCR system for detecting E. coli O157 from food sources. This slide shows the band for E. coli O157, using a kit we bought from the company Qualicon.
ABI 7700 (TaqMan) can be used for the detection of many agents:
ABI 7700 (TaqMan) is high throughput (96 samples per run). It is very fast; you can run 30-40 samples in 2 hours. It is quantitative, as it measures the viral or bacterial number, and the multiplex feature allows to target on different genes.
We do cost analysis tests comparing the cost of traditional methodologies vs. newer more rapid methods. For example, if the laboratory technician is testing for bordetella pertussis for six days, the tech labor time is much longer compared to the TaqMan, which takes only about two hours. TaqMan is useful for rapid answers in critical situations, outbreak investigations, and detection of non-cultivable and slow growing organisms. These tests are very useful in public health settings, but careful cost analysis is very important. The cost of the test using these rapid machines may be high but in the long run, the rapid tests can reduce the number of ill persons and major outbreaks.
As with molecular diagnostics, molecular epidemiology is the use of principles and methods of molecular biology in epidemiologic investigations of infectious diseases, for the purpose of identification of the outbreak/epidemic clone and tracking the clone to its source. It is also called strain typing: the process of analyzing multiple isolates within a given species to determine whether they represent a single strain or multiple strains.
Such strain typing systems are very useful in public health and clinical settings for epidemiological investigations. For example, if you have an E. coli O157 sample, they all look similar on culture; you cannot tell if the culture from patient A is different from the culture from patient B. Using molecular epidemiology, the strains can be identified. Thus strain typing can distinguish the real outbreaks from the non-outbreak strains. Strain typing can also define the pathogens of hospital acquired infection (endogenous vs. exogenous acquisition). Strain typing optimizes management of the individual patients—it clarifies colonization or contamination vs. true infections, and distinguishes relapse from reinfection with the same organism.
There are many different technologies for molecular epidemiology, but what is being used increasingly in public health laboratories and the CDC is pulsed-field gel electrophoresis (PFGE) because the intra- and inter-laboratory reproducibility is very high. It is a nearly universal typing system, applicable for all bacterial typing. The analysis is very simple, and it has high discriminatory power for epidemiologically related strains.
Here is an example of the use of PFGE.
We have 3 patients, Patient 1, 2 and 3. We recovered two isolates from Patient 1 (A, B), two isolates from Patient 2 (C, D) and two isolates from Patient 3 (E, F). When these isolates are run on PFGE, the bacteria are trapped in a matrix in the agarose gel. After enzyme is added to the exposed DNA, the DNA is digested into several pieces. The segmented DNA pieces can now be put on a gel, where large pieces of DNA are segregated. After the gel is run, you compare the different lengths of the DNA. You can see that the bands in samples A and B are identical. In samples C and D, you see that bands 7 and 9 are different. In E and F, all bands are different.
If someone becomes ill from tainted food, he or she would first go see the doctor, who would send the stool culture to a laboratory, where the organism is identified .
The isolate is sent to the PHL, and PFGE analysis is conducted. Meanwhile, epidemiologists begin their investigation, and if the PFGE patterns match, epidemiologic investigation continues, and if the patterns differ, the epidemiologists would stop and investigate other illnesses.
If the PFGE patterns do match interviews would be conducted to see if there is a familial link (food-handling issue) or a common food source link (commercial distribution issue). If there is a familial link, there would be teaching and the investigation would be over. If there was a common food source link, the pattern would be posted on PulseNet (to allow all 50 states and CDC to view pattern), and states can electronically share their DNA fingerprints on-line.
There was a large multistate E. coli O157:H7 outbreak in 1993. In late 1992, a contaminated lot of hamburgers was distributed to fast food restaurants in the western states. There were over 700 cases, and in Washington State, there were 501 cases, 302 bloody diarrhea cases, 151 hospitalizations, 45 HUS complications, and 3 deaths. At the time, PFGE was not available at the state level, but PFGE done at CDC later verified a common E. coli O157:H7 strain. CDC used PFGE technology to distinguish pre-outbreak, outbreak and post-outbreak samples, and recognized the usefulness of PFGE. In 1996 - 1997, they put together the four PHL’s from Washington, Texas, Minnesota and Massachusetts along with CDC to develop a network that would perform DNA-based fingerprinting.
The network later became "PulseNet" which allows the electronic sharing of DNA fingerprints of food-borne bacteria, among states and with the CDC. By performing PFGE on foodborne bacteria from patients (and food when possible) and sharing their DNA fingerprints with other PulseNet laboratories and with the CDC, multi-state outbreaks of food-borne diseases can be detected early, preventing illness and saving lives.
In 1999, there was an orange juice-smoothie outbreak. Children’s Hospital in Seattle found three cases of salmonella serotype Muenchen on Saturday, June 19, 1999, and the PH lab received the samples Monday, June 21, 1999. Within one day, using our new DNA fingerprinting system, we found they were all matching, and on June 25, there were an additional five cases, and they also matched the PFGE pattern. One of the five samples came from an Oregon household; thus we knew this outbreak was also happening in Oregon. At the same time, our environmental section collected samples from the blenders used to prepare these drinks, and on the June 26, we found the isolates to have the same pattern as that of the patients. Finally, the orange juice was cultured, and found to have the same PFGE pattern; even though they were washing the blender, the pathogen was coming from the orange juice. Within seven days, we solved the entire mystery working with epidemiology and environmental health staff.
Standard PFGE procedure for gram-negative organisms could take up to five days, but with the new rapid PFGE procedure, the results could be obtained in a single day. PFGE and PulseNet are tools that support epidemiological investigations conducted by state departments of health.
By comparing strain patterns within and between states, it is possible to:
PFGE and PulseNet assist in prevention. With PFGE data to support epidemiological findings, government agencies are able to respond more rapidly to food associated outbreaks. For example, a relatively few cases can implicate a product and result in large amounts of the product being recalled, preventing additional cases. Food processing and handling have been changed in certain industries based on outbreak investigations. Companies are now pasteurizing commercial fruit juices and screening beef for E. coli O157:H7 in order to prevent any cases. Government agencies are also increasing education for the public about safe food handling and cooking. Parents are learning not to give rare hamburgers or unpasteurized juice to children.
In spring of 2000, 47 cases of salmonella serotype Poona were reported in six states in the US. In spring 2001, there were 50 cases in five states, and in spring 2002, there were 58 cases in 14 states. In all three years, case control studies were performed, PFGE was done and cantaloupe was implicated. The PFGE patterns for 2000, 2001, and 2002 were all different.
The three outbreak patterns were different, but the cantaloupe had been imported during the same time period each year from the same region. Even with relatively few cases each year (average 52 cases in >50 million population), PFGE and PulseNet can pick up and identify common strains each year in a timely fashion. As a result, in October 2002 the United States Food and Drug Administration asked for a certification program for imported cantaloupe (certification program currently under development).
PulseNet started with just four states and now is present in all 50 states.
Washington public health laboratory handles this region, and we serve as a reference lab for Oregon, Idaho, Montana, California and Hawaii. PulseNet has gone international and many countries are considering bringing the system into their countries to prevent food-borne outbreaks.
Microarray technology is the next step in our laboratory in DNA technology.
Right now serotyping salmonella can take up to eight days, and we have to send those salmonella samples to CDC for serotyping. This microarray system would be very fast and useful. This is called FISH fluorescence in situ hybridization, and is very cost-efficient.
This is an example of Mycobacterium tuberculosis. Right now, we use a kit which costs us $1500 to process about 30-40 samples, and with the FISH technology the cost would be less than $10 and more sensitive and specific than this kit.
This concludes a brief overview of Washington State Public Health Laboratories. PHL’s are essential for reference services, disease identification, outbreak investigations, laboratory method development and improvements, and emergency preparedness and response. Thanks to molecular epidemiology, PHL’s are becoming icreasingly important in outbreak investigations. New and rapid methodologies, such as real-time PCR and PFGE, are being used more often. In outbreak investigations, collaboration and communication with other agencies and sites are imperative, and the electronic sharing of data has become standardized through PulseNet.