By Dr. David Ayoub

The Rationale for TB Screening of Healthcare Workers (HCWs) and Other Low-risk Populations:

A Critical Review of CDC Policy


The Emperor Has No Clothes, Cough or Fever


1) Healthcare workers (HCWs) are identified by the CDC as a high risk group for development of TB; however, no current clinical data exists that supports that contention. The overwhelming majority of TB, as with most infectious diseases, occur in individuals with compromised immune systems.

2) Targeted TB testing in HCWs is only recommended by the CDC; however, local facilities have often initiated mandatory testing policies amongst employees, subject to employment termination for refusal.

3) According to the CDC, initiation of chemoprophylaxis in the TB positive HCW is not mandatory in circumstances of negative health exam, negative chest radiograph and absence of additional risk factors. The overwhelming majority of HCWs who test positive have normal clinical exams and radiographs.

4) Current randomized studies of chemoprophylaxis in the TB skin test positive, healthy HCWs do not exist. Some randomized studies in AIDS patients show TB disease occurs with higher incidence in those receiving therapy for latent TB compared with those receiving no therapy.

5) One analysis showed no benefit to treatment of LTBI in all non disease states despite risk factors. Risks associated with chemoprophylaxis for TB may outweigh potential benefits.

6) TB skin test is inaccurate but yet is considered the gold standard to diagnose infection. Because there is no better method to diagnose infection, its actual test accuracy is unknown.

7) Phenol, a component of Tubersol, is a highly toxic industrial chemical with numerous known health risks yet this is a component of Tubersol.

8) The CDC claims that the TB skin test is safe, yet the manufacturer states that NO known carcinogenicity studies have been performed. In fact, phenol, when injected intradermally, is associated with skin cancer development in test animals.

9) The CDC states that TB skin testing is safe in pregnant women yet manufacturers have NOT performed mutagenicity testing. This is alarming in consideration that phenol, a component of Tubersol, is a known mutagen.

Summary: TB skin testing should not be required of otherwise healthy HCWs unless safety and efficacy studies have proven a benefit in this low risk population. The FDA should halt the use of Tubersol pending the standard and usual safety testing has been performed, including carcinogenicity and mutagenicity testing.


The following document is a review of current TB screening policies for HCWs. This includes an analysis of two published documents. First, "The Core Curriculum on Tuberculosis" (4th edition, 2000), published by the U.S. Department of Health and Human Services and The Centers for Disease Control. The second, a joint statement published in the MMWR, June 2000, represents the works of the American Thoracic Society and the CDC.

After anti-TB medications became available in the 1940's, a gradual decline of the number of TB cases were reported from 1953 until about the mid 1980's when there was a 20% increase from 1985 through 1992 (1, p. 16). According to the CDC, the major factors for this rise were 1) a deterioration of TB public health infrastructure, 2) HIV/AIDS epidemic, 3) immigration and 4) transmissions in congregate settings. (It would seem most logical that, since HIV increases the risk of TB by as much as 100-fold, and AIDS was an entirely new disease entity coinciding with this period of TB resurgence, that HIV would be the most likely contributing factor for rising cases of TB.) The CDC claims that the deterioration of the TB public health infrastructure was a major factor for TB resurgence, yet, the CDC publication offers no supportive evidence of this conclusion. If this were true, there would be an increase in the incidence of TB amongst healthy HCWs. Data to this effect is absent. In fact, I have not discovered any published data that proves the hypothesis that the neglect in screening programs resulted in more cases of TB during this era.

It is interesting to note that the incidence of TB in the US has declined steadily since the 1900’s. This decline was noted in spite of the fact that pharmaceutical therapies were unavailable for nearly five decades. How did the incidence of TB decline in the absence of TB screening programs and chemoprophylaxis? The CDC’s contention that the small TB spike occurring in the late 1980’s was the result of deteriorating TB control infrastructure seems very questionable.

The unrealistic goals of the CDC

In 1989, the CDC announced the goal of eliminating tuberculosis from the US by 2010. Plans and task forces were then established to accomplish this goal. To apparently help achieve this goal, the CDC now concludes that healthcare workers are part of a "targeted" population of individuals who are at high risk for developing the infection (TB skin tested positive) and developing subsequent clinical disease of tuberculosis. Institutional TB skin testing is recommended for the staff of healthcare facilities (1, p. 25;90-91).

Elimination of TB is unachievable and unrealistic. First, our government’s open door immigration policy allows countless high risk individuals into the US undetected on a daily basis. How can those individuals be screened when our government refuses to identify illegal aliens and allows them access to the healthcare system? Secondly, since the majority of TB occurs in the immune compromised host, how will the disease be irradicated unless the coexisting conditions are eliminated. AIDS, cancer and chemotherapy populations grow each year. Thirdly, false negative skin testing alone will bypass significant numbers of infected individuals (even one case missed in screening is significant when the goals are 100% eradication and the CDC claims a 23% transmission rate!) .


Risk analysis for TB and the rationale to screen HCWs

There are an alleged 10-15 million infected (skin test positive) persons in the US (1, p20:no source given). Of these, if not detected and no preventative treatment is initiated, the CDC states that 10% will develop TB at some point, 5% within the first 1-2 years, in spite of normal immune system (1, page 7; (2) page 8). The primary source of this data is not referenced in the CDC publications. Accurate natural historical data is critically important in order to support screening of asymptomatic HCWs. A study recently published in JAMA (3) challenges the CDC report and showed that of the estimated worldwide TB infection (TB test positive) rate of 32%, only 7.96 million cases of disease were reported in 1997, or a TB disease incidence of less than 0.2% amongst infected individuals (assuming a 6 billion world population). This is far less than the 5-10% rates quoted in the CDC publications and are consistent with the general concept that TB is a disease of opportunity, generally harmless to the immune competent host. This data alone should challenge the wisdom in screening otherwise healthy populations.

In addition, the CDC quotes a transmission rate of 21-23% (ref 1, page 6): this seems alarmingly high (referenced from “CDC Program Management Report”-unavailable). This implies that 21-23% of all contacts with a TB patient will develop the infection or disease! Demographic data simply does not support this alarmingly high rate.

Summary of CDC’s High Risk Groups for Developing TB (modified from 1, p 8)



Substance abuse

Hematological and reticuloendothelial disease

Chronic malabsorption and malnutrition

Diabetes Mellitus

Prolonged steroid therapy

Solid organ transplantation

Cancer of head and neck

Chronic renal failure

Low body weight

Healthcare workers

Table 3 in Ref 2 (p 9) assigns relative risk values for many of these groups; however, missing in this table are relative risk data of HCWs with healthy immune systems!


Relative Risk



Diabetes mellitus


Chronic renal failure/hemodialysis




Jejunoileal bypass


Solid organ transplant





Carcinoma of head or neck


The HCWs receiving mandatory yearly testing should be informed of his relative risk to develop TB disease. With the sole exception of the HCWs, all individuals designated in the CDC publications as high risk are those with abnormal systemic or pulmonary immune defenses. But is this proven? Is it possible that, as with many other diseases, the integrity of the host immune response system is of far greater importance than the presence of mere exposure to microorganisms? In fact, if HCWs were not at higher risk than the general population, unless they had additional medical risk factors, screening of HCWs would be no more valid than screening 100% of the population.

Although historical studies showed higher infection and disease rates in HCWs in an era when the prevalence of TB was higher, modern era data suggests this is no longer the case. McKenna, et al (4) concluded that the “overall case rate of tuberculosis in healthcare workers was slightly lower than the natural rate....most healthcare workers do not appear to have a risk of clinically active tuberculosis greater than the general population”. This conclusion has been confirmed in other recent studies (5).


The TB skin test

Tubersol, manufactured by Aventis, is comprised of a purified protein derivative of the organism M. tuberculosis. Its efficacy as a screening test is derived from the delayed hypersensitivity response in the infected host after intradermal injection. The exact number of doses sold in the US annually is apparently a guarded secret (personal communication, Aventis Pasteur, Sept 12, 2003).

False negatives are thought to occur frequently. Listed causes (6) include anergy, recency of exposure, viral infections, various vaccinations, overwhelming infection, various drugs(steroids) and malignancies and any condition that can impair the cell mediated immune response (sarcoid, malnutrition) . False positives include nontuberculous infections and BCG vaccine state. In spite of these inaccuracies, the CDC states that for persons with latent TB infection who have a normal immune system, test sensitivity approaches 100% ( 2, p 11). This statement is ridiculous for several reasons. First, the TB skin test is the gold standard, so it is not possible to accurately gauge the incidence of false negative exams. The sensitivity of this test , in actuality, remains unknown. Secondly, false negative exams occur in the groups who are at the very highest risk for disease in the first place, meaning that the false negative tests weigh heavily against the efficacy of screening in the most important risk groups-the one's most likely to develop disease in the first place!

Compounding the inaccuracies of the TB skin test is the revelation that only one in three positive reactions are correctly classified as positive by screen test interpreters (7).

Serious untoward reactions to the Tuberculin substance have been reported. Adverse reactions include local skin reactions (vesicles, ulcers, necrosis, scarring), rashes, and anaphylaxis.

Shockingly, in spite of its widespread use, the manufacturer's insert 6) states that Tubersol has not been evaluated for its carcinogenic or mutagenic potentials or influence on fertility. This is surprising considering the widespread use and frequent repetition intervals of administration, particularly to the long term HCWs employee. It is also irresponsible for the CDC to state that tuberculin is safe and reliable throughout the course of pregnancy (1, p. 29). This is in direct opposition with the manufacturers statement that Tubersol is NOT tested for mutagenicity. It is a fact that a declaration of safety without testing is a declaration of assumed safety, not a proven scientific fact.

What can we gather from the toxicology of its components? Tubersol contains 0.28% phenol(5), which is known to be highly toxic to humans (8). The 1969 American Heritage Dictionary defines phenol as a “caustic, poisonous, white, crystalline compound...derived from benzene and used in various resins, plastics, disinfectants, and pharmaceuticals. Phenol is also known as 'carbolic acid.'” Amongst the known adverse reactions to phenol are:

-irritating to skin, eyes, mucous membranes in humans

-ingestion in humans may cause death, paralysis, weakness, seizures, coma, respiratory collapse

-animal testing has shown severe toxicity

-limited data available on the chronic effects in humans, but in humans has caused dermal inflammation and necrosis, arrhythmia's, hepatic enlargement and dysfunction.

-animal studies have shown chronic exposure effects the CNS, respiratory, renal and cardiovascular systems

-no human development and reproduction studies have been performed BUT...

-animal studies have shown reduced weight, growth retardation, abnormal development, increased maternal mortality and decreased maternal weight gain.

-no studies have been done in humans with regards to carcinogenicity BUT...

- animal studies show phenol applied to skin is a skin carcinogen in mice!

These findings should be embarrassing to the FDA and shocking to recipients of the TB skin test. The CDC has no supportive data to state unequivocally that this test agent “is both safe and reliable throughout the course of pregnancy” (1, p 29) when animal studies exist to the contrary and demonstrate that one of its constituents is a skin carcinogen! How did the FDA approve this agent for use in the tuberculin skin test? Without testing, no conclusions can be made as to the safety of Tubersol, regardless of what comments critics might offer such as, for example, "....but it is such a small dose". Has Aventis proven that Tubersol is safe? The FDA, CDC and Aventis simply cannot answer that question with available scientific data.

The myth of screening and prophylactic therapy for the skin test positive HCWs- is there any proof of benefit?

With regards HCWs, a 1992 survey of 210 hospitals in the U.S. calculates the tuberculin reactivity rate of only 0.64% (9). This rate would even be expected to be lower today with falling prevalence of TB. Is it really worth screening 156 health HCWs in order to discover one positive convertor? In turn that convertor invariably will have a negative clinical exam and radiograph.

Cost benefit analysis was studied at a time when the tuberculin agent was $10. The costs of screening was $4,500 per person eligible for treatment and and $350,000 per case of TB prevented (10). The current cost of tuberculin is $18. It is particularly disappointing that as few as 25% of treated individuals actually are able to complete therapy.

Do local TB statistics support screening studies? Illinois Department of Public Health Statistics (11) compiling TB disease in all counties between 1990-2001 reveals that 86% of 104 counties reported on average fewer than two TB cases per year and 67% as few as one case annually! How can screening healthy HCWs in those counties with such low disease prevalence be justified? It simply cannot.

Inconsistencies within the CDC guidelines are easily found. First, the CDC publications do not support blanket chemoprophylaxis for all TB skin positive individuals with normal health exams and radiographs, yet one wonders how the CDC could NOT suggest therapy if the CDC believes their own statistic that 10% of these individuals really were destined to developed disease. If a subsequent TB positive test results in a negative clinical exam and chest radiograph and a decision NOT to treat, why not replace the risky, inaccurate TB test with a clinical exam and an employee radiograph? The decision to test SHOULD be a decision to treat.

Secondly, in order to justify chemoprophylaxis, outcomes studies must show proven safety, efficacy and long term benefit. If there were no proven benefits to the treatment group over non treated individuals, then the screening program would be without merit. Are there studies that support better outcomes in the treated group vs. untreated group in healthy HCWs? This author has discovered no such literature. Other studies challenge the supposition of beneficial chemoprophylaxis.

In Ref 2 (pg. 12), there is an allusion to seven trials in the 1950’s -60’s which demonstrated reduction of TB disease in the 25-92% range with chemoprophylaxis. Since these studies were done in the decades when TB was significantly more prevalent, approximately 50 years ago, they are no longer valid. Secondly, these studies almost exclusively involved non-US participants in countries where environmental and health issues were substantially different than in the US. This is not applicable to the current issue of treatment outcomes in healthy US HCWs treated for LTBI. The MMWR report (2, p 13) also refers to the IUAT trial which indicated a reduction of LTBI by 65-75%. Unfortunately, this trial was also performed in non-US individuals, all of whom had abnormal chest radiographs. This is a significantly different population than the typical US HCW who rarely displays an abnormal radiograph (2, pg. 11, and personal experience). The MMWR report attempts to further inflate these success statistics by quoting a 69-93% efficacy in “compliant” participants, a statistic that has little significance in real-life clinical outcomes. A closer look at the success rates shows they are quoted in meaningless percentages (if i bought three lottery tickets instead of one, does the 200% improved chance of winning mean anything in the real world?) Where is the statistical proof (for CDC authors, that would be the “p” values)? For the group with fibrotic lesions < 2 cm (the group closest to the most typical HCWs), there was NO STATISTICALLY SIGNIFICANT DIFFERENCE in placebo treatment and 12 week, 24 week and 52 week regimens. In addition, these trials were conducted from 1969-1977, over 30 years ago when TB rates were generally higher than today.

The MMWR report reviewed randomized treatment of LTBI in HIV positive individuals in seven studies from 1980-1997, although only one was strictly in the US (2,p 16). These results were decisively mixed, with some studies actually demonstrating higher TB rates in groups receiving isoniazid than nontreated groups! If benefit of prophylaxis is at best equivocal in high risk individuals, on what basis does the CDC use to justify prophylaxis in the healthy HCWs?

In a more recent study involving a New York area population of HIV-positive patients with anergy, there was a no significant difference in TB rates in those receiving anti-TB prophylaxis with INH compared with placebo (12). Once again, if efficacy of therapy for LTBI in the highest risk group is unproved, how can we justify prophylaxis in low risk HCWs with an intact immune system?

In order to answer these questions more completely with regards to the low risk HCWs, Tsetat et. al. (13) performed a decision analysis of screening test positive adults and concluded that “from the perspective of the individual adult with a positive skin test for TB, we cannot make a case for INH therapy”. Furthermore these authors distinguished TB mortality rates from individuals dying with their disease rather than because of their disease and calculated lower TB mortality rates than earlier authors. With this factored into the risk-benefit analysis they further concluded that “it does not matter how high the rate of developing TB is-the preferred strategy is always “NO INH”. These authors do not recommend therapy unless active TB disease is detected.

The FDA safety record with anti-tuberculous drugs

The 2000 MMWR report states:

“In 1965, when Isoniazid was first recommended in the United States for treatment of LTBI , it was not thought to cause severe toxicity. However studies in the late 1960’s suggested that isoniazid did cause hepatitis....It was not until the 1970’s that when several persons receiving isoniazid for LTBI died from hepatitis that the likelihood of isoniazid hepatitis was understood” (2, p 15-16).

Most clinicians are now very aware of the dangers of INH therapy and the need for careful clinical evaluation of all patients receiving therapy for LTBI. Unfortunately, this failure of the CDC/FDA to detect toxicity in recommended drug therapies prior to their widespread use and acceptance was not an isolated incidence.

The CDC manual published in 2000 (1, pg. 78) did not discuss potential hepatotoxicity of two additional anti-TB drugs, rifampin and pyrazinamide. In 2003, the CDC published a retraction of the recommendation of these drugs for LTBI based on the discovery of 48 cases of severe liver injury and 11 related deaths (14). An alarming 5% of patients who started this therapy did not complete the regimen due to hepatic toxicity. With these high percentages of complications it is difficult to comprehend how such severe adverse reaction rates were not discovered in pilot studies before the CDC issued widespread recommendations for their use in LTBI. Understandably, the FDA’s and CDC’s track record in TB drug safety testing would leave one even more concerned of their widely held opinion of the untested but assumed safety of the phenol containing skin testing agent Tubersol.

When recommendations become mandatory

According to the CDC, the “risk assessment should identify which HCWs have the potential for exposure and the frequency with which the exposure may occur. This information can then be used to determine which HCWs to include in the skin testing program and the frequency with which they should be tested” (1, p 91). This site or occupation-specific risk assessment of all HCWs is a targeted testing program. It is uncertain how individual institutions implement targeted testing. Radiology technicians currently undergo mandatory yearly testing at Memorial Medical Center and Springfield Clinic. It is my understanding that employment can be terminated in HCWs refusing to be tested.

The FDA and CDC: conflicts of interest

The following was published in the Washington Free Press as the results of an UPI investigation(15).

In the year 2000, the U.S. House of Representatives Committee on Government Reform held hearings to examine conflicts of interest in the two official panels that control vaccine policy in the U.S. (there is one panel at the Centers for Disease Control and one at the FDA). Among the committee’s findings were widespread conflicts of interest among panel members in the form of financial ties to pharmaceutical companies who manufacture vaccines that the panels oversee. Following is a summary of the committee findings, assembled by Dr Joseph Mercola.

* The CDC routinely grants waivers from conflict of interest rules to every member of its advisory committee.

* CDC advisory committee members who are not allowed to vote on certain recommendations due to financial conflicts of interest are allowed to actively participate in committee deliberations and advocate specific positions.

* The chairman of the CDC’s advisory committee until recently owned 600 shares of stock in Merck, a pharmaceutical company with an active vaccine division.

* Members of the CDC’s advisory committee often leave key details out of their financial disclosure statements, and are not required to provide the missing information by CDC ethics officials.

* Three out of the five FDA advisory committee members who voted to approve the rotavirus vaccine in December 1997 had financial ties to the pharmaceutical companies that were developing different versions of the vaccine. The vaccine was recalled a few years later after numerous public complaints of serious bowel obstruction due to the vaccine.

* Four out of the eight CDC advisory committee members who voted to approve guidelines for the rotavirus vaccine in June 1998 had similar financial ties.

In a USA Today report of conflicts of interest on the 18 advisory committees established by the FDA, the following was reported (16).

* 54% of the experts hired to advise the government on safety and efficacy policies had financial relationships with the pharmaceutical companies that would be directly affected by their opinions

* since 1988, the FDA has waived on more than 800 occasions the federal law that would have other wise prohibited use of experts with financial conflicts

* 92% of FDA advisory meetings had at least one member with a conflict of interest

* 55% of FDA advisory meetings were held when at least half of the committee members had conflicts

* in 102 FDA advisory meetings dealing with the fate of a specific drug, 33% of the experts had a financial conflict.

The pharmaceutical industry enjoys the benefits of increased revenue when government regulatory bodies pass favorable legislature promoting use of its products. The influential power of this industry has been subject to much criticism.

According to Public Citizen’s report (17)

Drug industry lobbying ranks include 26 former members of Congress. All told, 342 lobbyists (51 percent of those employed by the industry) have "revolving door" connections between K Street and the federal government.

The Pharmaceutical Research & Manufacturers of America (PhRMA), which represents more than 100 brand-name prescription drug companies, shelled out $14.3 million last year, a 26 percent increase from 2001 and nearly double what the group spent in 2000. PhRMA hired 112 lobbyists in 2002, 30 more than the year before.

Brand-name drug manufacturers spent more than 20 times as much on lobbying as generic drug-makers – $76 million versus $3.4 million. And they employed seven lobbyists for every one hired by their generic counterparts. Biotechnology companies spent $12 million on lobbying.

Since Public Citizen began tracking the drug industry’s lobbying activities in 1997, the industry has spent nearly $478 million lobbying the federal government. In that same period, the top 25 pharmaceutical companies and trade groups gave $48.6 million to federal campaigns. Well over $100 million more went to paying for issue ads, hiring academics, funding non profits and other activities to promote the industry’s agenda in Washington. All told, the drug industry has spent nearly $650 million on political influence since 1997.

Drug company profits have been staggering.

*By comparison, all companies in the Fortune 500 suffered a combined loss of 66.3 percent in profits from 2001 to 2002. The pharmaceutical industry soared past other business sectors – raking in profits five-and-a-half times greater than the median for all industries represented in the Fortune 500.

*17% profit (as a percent of revenue) far outpaces the 3.1% median value for all other Fortune 500 industries.

*Profits registered by the 10 drug companies on the list were equal to more than half the $69.6 billion in profits netted by the entire roster of Fortune 500 companies – when all losses are subtracted from all gains.

The dollars available to the drug companies for influencing industry agencies are staggering. There is no question that the CDC’s policy of widespread use of Tubersol for testing HCWs has widely increased the market for their product. National advisory committees have been an essential and necessary part of healthcare policy in this country but has been also linked to significant conflicts of interest, as reported in JAMA (18). In a review of of doctors involved in establishing national guidelines on disease treatment, they found that :

85% of guideline authors have some sort of relationships with drug companies, and they are often not disclosed

38% of respondents said they had served as employees or consultants for drug companies; 58% received research money

59% had links with drug companies whose medications were considered in the particular guidelines they authored, almost all cases predating the guideline creation process

These numbers may be even greater, as only 52% of authors responded

These are disturbing revelations. Questions must be asked regarding the establishment of national TB skin testing policies for healthy HCWs. First, to what degree has Aventis benefited from the expansion of mandatory testing to healthy HCWs in the United States? Secondly, did advisory committee members who established TB skin testing policies have financial ties with Aventis?


There is no clinical scientific evidence that the healthy HCW is at higher risk than the general population. Furthermore, even in positive TB reactors, there is no modern scientific evidence that would support benefit of chemoprophylaxis for LTBI in healthy HCWs. The Tubersol agent in use has not been adequately tested for safety and its accuracy is questionable and unprovable. The TB skin testing policy for LTBI in the typical HCW is of doubtful efficacy and benefit and of unknown risk to the individual HCW. Mandatory testing is unsupported. Most tragically, our government health agencies have a tract record of errors in drug safety testing and these same agencies have conflicts of interest that raise serious questions of the mechanisms that healthcare policies are established.


Immediate suspension of mandatory TB skin testing policies of HCWs is reasonable considering the lack of clinical efficacy of testing or subsequent treatment of LTBI. A review of current local and state public health records could be undertaken to determine a) the incidence of TB skin test reactivity amongst HCWs, b) the true risk of TB disease in the skin reactive HCWs who fails to receive prophylaxis for LTBI and c) the outcome analysis of the HCWs placed on preventative drug therapy and finally d) the review or performance of pertinent toxicology studies on Tubersol that establishes this agent as safe.


1) Core Curriculum on Tuberculosis. What Every Clinician Should Know.

U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, Fourth Ed. 2000

2) Targeted tuberculin testing and treatment of latent TB infection. MMWR; 49, June 9, 2000

3)JAMA 1999;282:677-686

4) McKenna, MT, et. al., The association between occupation and tuberculosis. A population-based study. Am. J. Resp. Crit Care Med.154: 587-93, 1996.

5) Kwan, SYL, et. al., Nosocomial tuberculosis in hospital staff in a Hong Kong chest hospital. Chinese Med J. 103, 909-914, 1990.

6) Tubersol PDR, 2003

7) Underreading of the tuberculin skin test reaction. Kendig,, Chest,113,1175,1998


9) Fridkin, SK, et. al., SHEA-CDC Tuberculosis survey, Part 1. Status of tuberculosis infection control programs at member hospitals. Infect. Control Hosp Epidemiol. 16: 129-134, 1992

10) Raad, I. et. al., Annual tuberculin skin testing of employees at a university hospital: a cost-benefit analysis. Inf. Control Hosp Epidemiol; 10, 465-9, 1989.


12) F. Gordin, et. al., NEJM; 337(5), 315-320, 1997.

13) Tsevat, J, et. al., Isoniazid for the tuberculin reactor: take it or leave it. Am Rev. Respir. Dis., 137: 215-220, 1988.

14) Update: Adverse event data and revised American Thoracic Society/CDC recommendations against the use of rifampin and pyrazinamide for treatment of latent tuberculosis infection---United States, 2003. MMWR; 52(31), 735-39, 2003.



17) margins by drug companies

18) N.K. Choudhry, et. al. Relationships between authors of clinical practice guidelines and the pharmaceutical industry JAMA; 287,612-617, 2002.