Typhoid Fever in the United States, 1999-2006:   FREE

Infection with Salmonella ser Typhi causes an estimated 20 million cases of typhoid fever and 200 000 deaths annually worldwide.¹ In the United States, typhoid fever is now a rare disease, with approximately 300 clinical cases reported per year.² Dramatic declines in incidence of and mortality due to typhoid fever were observed in the United States after widespread implementation of municipal water and sewage treatment systems in the first half of the 20th century.³ In recent years, the majority of cases in the United States have been associated with foreign travel.^4- 7

Although the risk of typhoid fever may be reduced with attention to water quality, food hygiene, and vaccination, effective treatment of S Typhi infection is needed when these measures fail. Over the last 20 years, emergence of S Typhi strains resistant to antimicrobial agents has complicated treatment of infected patients.⁸ Resistance to antimicrobial agents has been documented from many high-incidence areas and among isolates from patients diagnosed and treated in the United States. Multidrug-resistant strains of S Typhi [MDRST] that exhibit resistance to the commonly used first-line antimicrobial agents ampicillin, chloramphenicol, and trimethoprim-sulfamethoxazole emerged during the 1980s, predominantly from South Asia and travelers returning from that region.^9- 10 Recent estimates of the proportion of multidrug-resistant infections among US patients range from 12% during 1985-1994⁵ to 17% during 1996-1997.⁶ In addition, identification of nalidixic acid–resistant S Typhi (NARST) and reports of infection with S Typhi strains resistant to ciprofloxacin from typhoid-endemic areas have generated concern that strains resistant to fluoroquinolones may become more prevalent.¹¹ Rates of nalidixic acid resistance are of particular concern. Although this older antibacterial is rarely used for treatment, resistance to nalidixic acid can be a marker for decreased susceptibility to fluoroquinolones.

Surveillance for typhoid fever cases and resistance among S Typhi strains isolated from patients diagnosed and treated in the United States can be used to inform treatment recommendations and provide guidance for US travelers to typhoid-endemic countries. Since most US patients acquire their infections abroad, these cases also provide insight into global patterns of S Typhi antimicrobial susceptibility.

Detailed epidemiologic information on laboratory-confirmed cases of typhoid fever in the United States is collected from state and local health departments through the National Typhoid Fever Surveillance System at the Centers for Disease Control and Prevention (CDC). Before 1999, information in typhoid fever surveillance reports regarding antimicrobial susceptibility of S Typhi isolates was limited by incomplete reporting and nonuniform testing methods among multiple reporting laboratories. In 1999, national laboratory-based surveillance for antimicrobial resistance of S Typhi isolates was initiated through the National Antimicrobial Resistance Monitoring System for Enteric Bacteria (NARMS). We linked ongoing national typhoid fever case surveillance and systematic laboratory surveillance for antimicrobial resistance of S Typhi isolates to assess trends in risk factors and resistance patterns among US typhoid fever cases from 1999 through 2006.

Since 1975, the CDC has received surveillance case reports from state and local health officials on bacteriologically confirmed cases of typhoid fever. A case of typhoid fever is defined as an acute illness compatible with typhoid fever and S Typhi isolated from a sterile site, stool specimen, or urine culture. Patient demographic and clinical information, typhoid vaccination, and travel history are collected by state and local health department officials using a standard form (CDC form 52.5). Travel-associated typhoid fever is defined as a case of typhoid fever meeting the national surveillance case definition in a person who traveled outside of the United States in the 30 days before illness. Countries visited by travelers were categorized by continental and subcontinental regions according to United Nations criteria. Domestically acquired typhoid fever was defined as typhoid fever meeting the national surveillance case definition in persons who had not traveled outside the United States in the 30 days before onset of symptoms. We reviewed reports of outbreaks of typhoid fever from the CDC Foodborne Outbreak Reporting System for 1999-2006.

Information on the number of US residents traveling abroad and nonresident foreign visitors traveling to the United States for 1999-2006 was obtained from Tourism Industries, International Trade Administration, US Department of Commerce (http://tinet.ita.doc.gov/about/index.html) and used to estimate rates of typhoid fever among travelers to specific regions and countries. The rate of travel-associated typhoid fever was calculated using the number of nonimmigrant travel-associated cases each year divided by the total number of US resident travelers and foreign visitors.

The NARMS laboratory at the CDC receives isolates from state and local public health laboratories for antimicrobial susceptibility testing. NARMS had 17 participating laboratories in states and localities comprising 40% of the US population during 1999-2001, 28 laboratories in 2002, and beginning in 2003, NARMS was nationwide in 54 laboratories. Since 1999, NARMS has received all S Typhi isolates from participating laboratories for antimicrobial susceptibility testing. During 1999-2006, S Typhi isolates were tested in the NARMS laboratory by broth microdilution for susceptibility to at least 14 antimicrobial agents: amikacin, ampicillin, amoxicillin–clavulanic acid, ceftiofur, ceftriaxone, chloramphenicol, ciprofloxacin, gentamicin, kanamycin, nalidixic acid, streptomycin, sulfamethoxazole/sulfisoxazole, tetracycline, and trimethoprim-sulfamethoxazole. (Additionally, isolates were tested to cephalothin during 1999-2003 and to cefoxitin during 2000-2006.) Clinical and Laboratory Standards Institute minimum inhibitory concentration (MIC) interpretive standards for Enterobacteriaceae were used for resistance thresholds for susceptibility testing.¹² Multidrug-resistant S Typhi isolates were defined as those resistant to ampicillin, chloramphenicol, and trimethoprim-sulfamethoxazole. Nalidixic acid resistance was defined as an MIC of 32 μg/mL or more and resistance to ciprofloxacin was defined as 4 μg/mL or more. Decreased susceptibility to ciprofloxacin was defined as an MIC of 0.125 μg/mL or more. Susceptible strains were sensitive to ampicillin, chloramphenicol, trimethoprim-sulfamethoxazole, and nalidixic acid.

Data from surveillance case reports on US typhoid cases and laboratory data from NARMS were linked by comparing available identifying information: state public health laboratory identification number (when available) or state, year of diagnosis, first 3 letters of surname, date of birth, age, and sex. We reviewed the combined data to assess risk factors for S Typhi infection and antimicrobial resistance of S Typhi isolates. Statistical analyses were conducted using SAS software, version 9 (SAS Institute Inc, Cary, North Carolina). We calculated odds ratios and 95% confidence intervals (CIs) for categorical variables and compared differences in medians using the Wilcoxon 2-sample test. Trends were assessed using the Cochran-Armitage trend test. P values were 2-tailed and considered significant at P<.05.

For the years 1999-2006, 1902 surveillance case reports were received at the CDC and 2016 isolates were received and tested by the NARMS laboratory (Table 1). We identified 1131 cases with epidemiologic information from case reports and antimicrobial testing information from the NARMS laboratory. The proportion of typhoid fever surveillance reports received from NARMS participating states for which an isolate was identified in the NARMS laboratory increased from 41% (95% CI, 34%-48%) in 1999 to 71% (95% CI, 67%-75%) in 2005. Since 2003, when NARMS received isolates from all 50 states, more than 65% of surveillance reports could be matched to a NARMS isolate each year.

During 1999-2006, surveillance reports were received from 46 states on 1902 typhoid fever cases. More than half of the cases were reported from California, New York, and New Jersey (Table 2). The median age of typhoid patients was 22 years, with a range of younger than 1 year to 90 years. Of 1685 patients whose age was known, 64 (4%; 95% CI, 3%-5%) were younger than 2 years, 205 (12%; 95% CI, 11%-14%) were 2 to 5 years old, and 423 (25%; 95% CI, 23%-27%) were 6 to 17 years old. Eight hundred seventy-four patients (46%; 95% CI, 44%-48%) were female. Among 1765 patients with clinical information, 1295 (73%; 95% CI, 71%-75%) were hospitalized for a median duration of 6 days (range, 1-25 days). Three patients died (case-fatality rate, 0.2%; 95% CI, 0%-0.5%) Two of these patients were travelers who had arrived from India; 1 had immigrated from Mexico.

Of 1830 typhoid cases with travel information, 1439 (79%; 95% CI, 77%-81%) were travel-associated. The proportion of travel-associated cases increased from 64% (95% CI, 62%-66%) in 1999 to 85% (95% CI, 83%-87%) in 2006. Most travelers, 952 (66%; 95% CI, 64%-68%) of 1439, reported that their reason for travel was visiting friends and family; 10% (95% CI, 8%-12%) were immigrants; 9% (95% CI, 8%-11%) were tourists; 3% (95% CI, 2%-4%) traveled for business; and 6% (95% CI, 5%-7%) reported other reasons for travel. Only 58 (5%; 95% CI, 4%-7%) of 1094 travelers who reported vaccination status had received any typhoid vaccine within 5 years of their trip. Of the 25 travelers who reported the type of vaccine received, 5 had received the old parenteral typhoid vaccine (Wyeth-Ayerst, Collegeville, Pennsylvania), 13 had received Typhim Vi (Sanofi Pasteur, Lyon, France), and 7 had received Vivotif (Berna Biotech, Bern, Switzerland) oral typhoid vaccines.

Among 1277 travelers with typhoid fever who reported visiting a single country, 78% (95% CI, 76%-80%) had traveled to a country in Asia, while 17% (95% CI, 15%-19%) had traveled to Mexico, Central America, South America, or the Caribbean and 4% (95% CI, 3%-5%) had traveled to Africa. Three countries on the Indian subcontinent, India, Pakistan, and Bangladesh, were visited by 67% (95% CI, 64%-70%) of patients with travel-associated typhoid cases (Table 3).

The overall rate of travel-associated typhoid fever during 1999-2006 was 1.6 per 1 million travelers arriving in the United States. The typhoid fever rate for travelers arriving from countries other than Canada (from which no typhoid patients had traveled) was 2.2 per 1 million travelers. For travelers arriving from Mexico, Central America, South America, or the Caribbean, the typhoid fever rate was 1.3 per 1 million; for travelers arriving from Africa, the rate was 7.6 per 1 million; and from Asia, the rate was 10.5 per 1 million. Among countries for which travel data were available for the entire study period, India had the highest rate, with 89 typhoid cases per 1 million travelers. The rate for travel to India varied from 55 per 1 million travelers in 1999 to 122 per 1 million in 2003.

Among 391 domestically acquired cases, 54 (17%; 95% CI, 14%-22%) of 313 cases for which information was reported were traced to a typhoid carrier; 7 (13%; 95% CI, 4%-22%) of these carriers were previously known to the reporting health department. Seventy-five (22%; 95% CI, 18%-26%) of 346 domestically acquired cases for which information was reported were part of typhoid outbreaks. However, only 3 outbreaks with a total of 13 laboratory-confirmed cases were reported to the CDC Foodborne Outbreak Reporting System during this time: 1 outbreak in Texas (6 confirmed cases) due to contaminated oysters harvested from the US Gulf Coast¹³; 1 in Florida (3 cases) due to beverages prepared from frozen imported tropical fruit,¹⁴ and 1 in Maryland (4 cases) associated with a restaurant.

During 1999-2006, the NARMS laboratory received 2016 isolates for antimicrobial susceptibility testing from 46 states. Specimen source was identified for 1985 isolates tested: 1413 (71%; 95% CI, 69%-73%) were from blood, 460 (23%; 95% CI, 21%-25%) were from stool, 36 (2%; 95% CI, 1%-3%) from urine, and 76 (4%; 95% CI, 3%-5%) from other sterile sites.

Among 2016 isolates tested, 1155 isolates (57%; 95% CI, 55%-59%) were susceptible to all 14 agents tested, while 861 (43%; 95% CI, 41%-45%) were resistant to at least 1 antimicrobial agent. Multidrug resistance was found in 272 isolates (13%; 95% CI, 12%-15%); nalidixic acid resistance was found in 758 isolates (38%; 95% CI, 35%-40%). Two hundred eight isolates (10%; 95% CI, 9%-12%) were both MDRST and NARST. Among NARST isolates, 734 tested throughout the study period had decreased susceptibility to ciprofloxacin (MIC >0.12 μg/mL), representing 97% (95% CI, 96%-98%) of NARST and 36% (95% CI, 34%-38%) of all isolates tested. Of 770 total isolates with decreased susceptibility to ciprofloxacin, 36 (5%; 95% CI, 3%-6%) were not NARST. Five isolates (0.2%; 95% CI, 0.03%-0.5%) were resistant to ciprofloxacin (all MICs = 4 μg/mL) and 2 isolates were resistant to ceftriaxone (both MICs = 64 μg/mL).

The proportion of S Typhi isolates with multidrug resistance and nalidixic acid resistance varied by year (Figure 1). The proportion of all isolates that were MDRST ranged from a high of 19% (95% CI, 15%-23%) in 2001 and 2006 to a low of 6% (95% CI, 3%-9%) in 2002. The proportion of isolates that were NARST increased from 19% (95% CI, 13%-25%) in 1999 to 54% (95% CI, 49%-59%) in 2006 (P = .001).

A similar trend for NARST was seen for S Typhi isolates submitted by the 17 sites participating in NARMS throughout 1999-2006 (Figure 2).

We identified 1131 typhoid fever cases from 39 states with both a surveillance case report and associated antimicrobial susceptibility testing information on an S Typhi isolate from the same patient. The descriptive epidemiology and proportion of resistant isolates associated with these 1131 cases were similar to that of all typhoid cases. Patient median age was 21 years (range, <1-90 years) and 47% (95% CI, 44%-50%) were female. Among 1070 patients with reported clinical outcomes, 806 (75%; 95% CI, 73%-78%) were hospitalized and 1 (0.1%; 95% CI, 0%-0.3%) died. Travel outside the United States in the month before illness onset was reported by 869 (80%; 95% CI, 77%-82%) of 1088 patients. Of 1131 S Typhi isolates tested for these cases, 482 (43%; 95% CI, 40%-46%) were resistant to at least 1 antimicrobial agent; 142 (13%; 95% CI, 11%-14%) were MDRST and 440 (39%; 95% CI, 36%-42%) were NARST.

The median age of patients infected with MDRST, NARST, or susceptible strains was similar (Table 4). Patients with MDRST and NARST were no more likely to be hospitalized than patients with susceptible infections, and the median number of days hospitalized was the same. A higher proportion of MDRST patients (96%; 95% CI, 92%-99%) had traveled outside of the United States compared with patients who had susceptible infections (68%; 95% CI, 65%-72%). Among 110 patients with MDRST infections who reported travel to a single country, India (n=37 [34%; 95% CI, 23%-42%]), Bangladesh (n=30 [26%; 95% CI, 19%-36%]), Pakistan (n=27 [25%; 95% CI, 17%-33%]), and Cambodia (n=8 [7%; 95% CI, 2%-12%]) were most commonly named. Among 374 patients with NARST infections who reported travel to a single country, 361 (97%; 95% CI, 95%-98%) had traveled to these same 4 countries: 278 (74%; 95% CI, 70%-79%) traveled to India, 41 (11%; 95% CI, 8%-14%) to Bangladesh, 33 (9%; 95% CI, 6%-12%) to Pakistan, and 9 (2%; 95% CI, 1%-4%) to Cambodia. Infection with resistant strains of S Typhi was strongly associated with travel to 1 of 3 countries on the Indian subcontinent, India, Pakistan, or Bangladesh (MDRST odds ratio, 7.5; 95% CI, 4.1-13.8; NARST odds ratio, 20.4; 95% CI, 12.4-33.9).

The proportion of resistant infections among 864 travel-associated cases varied by the region to which patients had traveled. Among patients with typhoid fever who had traveled to the Indian subcontinent, 17% (95% CI, 15%-20%) were infected with MDRST and 65% (95% CI, 62%-68%) with NARST. The corresponding proportion of MDRST and NARST infections among patients who had traveled to Southeast Asia was 14% (95% CI, 12%-16%) and 20% (95% CI, 17%-23%); for Central or South America was 1% (95% CI, 0.3%-2%) and 3%(95% CI, 2%-4%); and for Africa was 24% (95% CI, 21%-27%) and zero. For individual countries, typhoid patients who had traveled to India had the highest rate of drug-resistant infections. The proportion of typhoid patients who had traveled to India and were infected with MDRST varied by year (Figure 3), whereas the proportion of patients infected with NARST after traveling to India increased steadily during this period, from 44% (95% CI, 12%-76%) in 1999 to 88% (95% CI, 82%-94%) in 2006. Among cases linked to a NARMS isolate, the proportion of travel-related typhoid cases associated with travel to India varied little during this time.

Limited epidemiologic information was available for the 5 cases of ciprofloxacin-resistant S Typhi infection. They were reported from 4 states, Arizona, California (n=2), Massachusetts, and Texas. One case occurred in 2003, 1 in 2005, and 3 in 2006. Patient ages ranged from 1 to 26 years; 4 patients were younger than 9 years. Four patients were hospitalized; none died. None of these patients were reported to be typhoid vaccine recipients, though 2 were younger than 2 years old and therefore below the recommended age for any available typhoid fever vaccine. All 5 patients had arrived from India; 4 reported visiting family and 1 was immigrating to the United States.

Although typhoid fever remains a rare disease in the United States, it continues to cause substantial morbidity among affected patients. Patients with typhoid fever during the period of this study were hospitalized in nearly three-quarters of reported cases, half for more than a week. The low case-fatality rate among reported typhoid fever cases in the United States (0.1%) compared with published reports from other regions^15- 17 may be due in part to greater access to supportive care and appropriate antibiotic therapy. However, proper treatment of infected patients with effective antibiotic therapy is increasingly complicated by strains resistant to available antimicrobial agents. During 1999-2006, resistance to antimicrobial agents was common among S Typhi isolates from patients with typhoid fever in the United States. Strains resistant to first-line agents have been noted for some time and the proportion of MDRST infections in this report (13%) is comparable with other recent estimates.^5- 6 This proportion of MDRST infections among typhoid patients is, in turn, comparable with susceptibility testing reports from typhoid-endemic areas.^9- 10 In a recent summary of typhoid disease burden in 5 Asian countries, the overall percentage of MDRST was 23%.¹⁸

Notably, about one-third of all S Typhi isolates in this study were resistant to the quinolone nalidixic acid, and the proportion of NARST strains among US patients steadily increased during the study period. During 1996-1997, enhanced surveillance revealed that 7% of S Typhi strains tested were resistant to nalidixic acid; by 2006, 54% of strains were NARST. This increase appears to be related to an increase in the proportion of NARST infections among patients with typhoid fever who traveled to southern Asia, rather than changes in reporting to NARMS or an increase in travel to southern Asia among all typhoid cases. Surveillance for antimicrobial resistance among S Typhi isolates conducted recently in this region yielded similarly high proportions of nalidixic acid resistance: 57% of S Typhi isolates from India and 59% of isolates from Pakistan were NARST.¹⁸

Although nalidixic acid is not commonly used for treatment of S Typhi infections, resistance to nalidixic acid is a marker for decreased susceptibility to fluoroquinolones. Nearly all NARST isolates in this study had decreased susceptibility to ciprofloxacin, a commonly used oral agent for S Typhi infection.⁸ In addition, a high proportion of isolates with decreased susceptibility to ciprofloxacin were NARST. However, since NARST may not identify all isolates with decreased susceptibility to ciprofloxacin, the MIC for clinically used fluoroquinolones should be measured when possible. Recent reports indicate NARST infection has been associated with poor clinical outcomes among US patients, including prolonged fever and a higher rate of treatment failure.¹⁹ Although decreased susceptibility to ciprofloxacin was common among the S Typhi isolates tested, infection with a ciprofloxacin-resistant S Typhi strain has, so far, been rare. The 5 cases of infection with ciprofloxacin-resistant strains are the first reported among typhoid fever patients treated in the United States and follow reports of ciprofloxacin-resistant cases from other countries.²⁰

These data indicate that recommendations for appropriate empirical therapy for typhoid fever patients should be made with caution. Resistance to ampicillin and to trimethoprim-sulfamethoxazole remains prevalent and precludes the use of these agents as initial therapy. Although nalidixic acid resistance and decreased susceptibility to ciprofloxacin is of concern, a fluoroquinolone remains an appropriate choice for empirical therapy in adults. Among children, in whom fluoroquinolones are limited to off-label use, third-generation cephalosporins are appropriate. Further therapy should be guided by close monitoring of patients' clinical response and antimicrobial susceptibility testing of S Typhi isolates at a clinical diagnostic laboratory.²¹ At the national level, resistance among S Typhi isolates will be monitored closely in NARMS for trends that may influence therapeutic recommendations.

Typhoid fever in the United States and infection with resistant S Typhi strains continues to be associated with foreign travel, specifically with travel to the Indian subcontinent. The proportion of US typhoid cases associated with foreign travel has increased steadily over the past several decades. During 1975-1984, 62% of typhoid fever cases were related to foreign travel⁴; during 1985-1994, 72% of cases were travel-related⁵; 74% in 1995-1999⁷; and 79% in this study. In 2006, foreign travel was reported by more than 85% of typhoid fever patients. An increase in the proportion of typhoid fever patients reporting travel to southern Asia and a concomitant decline in the proportion of typhoid patients reporting travel to Mexico has been an ongoing trend since the 1980s.⁵ In this study, we observed a higher rate of typhoid fever among travelers returning from Asia (10.5 per 1 million travelers) or Africa (7.6 per 1 million) than among those returning from Mexico, Central America, South America, or the Caribbean (1.3 per 1 million). Improved sanitation in response to the cholera epidemic in South America during the 1990s may have played a role in lower rates of endemic typhoid in this region and, consequently, lower rates of typhoid fever among travelers to that region.²²

With the threat of increased antimicrobial resistance, prevention remains paramount.²³ Travelers to the Indian subcontinent and those visiting friends and family are singularly important groups on which to focus enhanced outreach and typhoid fever prevention efforts.²⁴ Travelers to high-risk areas should ensure they are vaccinated against typhoid fever and take appropriate precautions regarding food and beverages. More than 96% of travel-related cases occurred in persons who were older than 2 years, an age group for which typhoid vaccine is available. Currently licensed typhoid vaccines, Vivotif and Typhim Vi, are reportedly 70% and 90% effective in preventing typhoid infection among residents of typhoid-endemic areas.⁸ Recent analyses of typhoid fever among US travelers indicate that typhoid vaccines are well-tolerated and vaccination should be considered even among travelers planning trips of less than 2 weeks' duration to typhoid-endemic areas.⁷ Because neither commercially available vaccine is 100% effective, food and water precautions remain important. Further advice on food and water safety while traveling can be found at http://www.cdc.gov/ncidod/dbmd/diseaseinfo/typhoidfever_g.htm. Currently available vaccines are not licensed in the United States for children younger than 2 years, so efforts aimed at reducing exposure to contaminated water and food are especially important for this age group.

This report has several limitations. We were able to link laboratory testing results to approximately 60% of typhoid fever cases reported to national surveillance during the study period. Linkage was limited by missing information and the small number of laboratories submitting isolates to NARMS early in the study period. However, similar descriptive epidemiology of typhoid cases linked to NARMS and of all cases, as well as a similar proportion of resistant isolates among those linked to cases and among all isolates tested, suggest that the linked cases and isolates are a reasonable representation of all reported typhoid cases and tested isolates. Rates of typhoid among travelers should be interpreted with caution. Counts of foreign travelers are estimates from custom forms and from surveys of a sample of travelers; complete travel information for the entire study period is not available for all countries.

Given trends in resistance for S Typhi, close monitoring of resistance patterns of isolates and characteristics of typhoid patients is critical. Ongoing antimicrobial testing of S Typhi isolates at a central laboratory linked to ongoing national typhoid case surveillance can help guide recommendations for US patients. Since most typhoid fever among patients treated in the United States is acquired abroad, ongoing surveillance may also help track global patterns of antimicrobial resistance of S Typhi. Reducing the burden of typhoid fever in the United States will require increased attention to prevention measures by travelers, including improved vaccination coverage among travelers to typhoid-endemic areas. Further reductions in typhoid fever among travelers will depend on increased availability of safe drinking water as well as improved sanitation and food hygiene in typhoid endemic areas, measures that would go a long way toward reducing the global burden of typhoid fever.²⁵