Following the introduction of potent new antiretroviral therapies, the French Ministry of Health convened, in April 1997, a working group of scientists, clinicians, public health experts, an ethicist and members of affected communities to discuss the issue of providing antiretroviral drugs for non-occupational HIV exposures . In the climate of optimism of patients and clinicians that followed these major therapeutic advances, the working group elaborated formal recommendations on post-exposure prophylaxis (PEP) for non-occupational HIV exposures. The decision to provide antiretroviral drugs for non-occupational HIV exposure was based on the results of the effectiveness of zidovudine in preventing perinatal transmission  and transmission among health care workers exposed to potentially infected blood , the evidence that PEP prevented retroviral infection in some animal studies [4,5], an ethical criterion of equal access to PEP for HIV exposures of a contamination risk similar to that of health care workers, results of a preliminary survey of physicians’ non-occupational PEP practices in France  and a position paper that proposed PEP as a possible means of secondary HIV prevention . However, following a meeting of expert consultants convened in July 1997 by the Centers for Disease Control and Prevention (CDC), the US Public Health Service could not ‘definatively recommend for or against antiretroviral drugs for non-occupational HIV exposures because of the lack of efficacy data’ on the use of such therapies for this type of exposure . In Europe, Switzerland was the first country to publish recommendations on non-occupational PEP in December 1997 .
French recommendations were issued in April 1998 . Requests for non-occupational PEP are dealt with in hospital emergency rooms on a 24 h a day basis. There, individuals are evaluated and given PEP if necessary. If a medication is prescribed, patients are required to consult within 48 h a ‘PEP reference physician’ (specialized in treating HIV exposures) to confirm the prescription. Information on access to PEP was included in leaflets on HIV testing and on specific leaflets targeting homosexual people and injecting drug users (IDU). In addition, centres caring for victims of rape incorporated PEP recommendations in their comprehensive rape treatment programme.
To evaluate the impact of the recommendations on access to PEP, trends of PEP request and prescription levels, practices and attitudes of physicians and to estimate the risk of side-effects of drugs used for non-occupational PEP, we carried out two surveys among physicians, the first in 1997 before any recommendations had been published  and the second in 1999, a year after their publication.
Materials and methods
The study population consisted of hospital physicians taking care of HIV-infected patients whatever their medical specialities. The number of eligible physicians was 1604 in the 1997 survey. The 1998 recommendations on non-occupational PEP emphasized the role of ‘PEP reference physicians'. These were appointed in every hospital with emergency rooms among physicians taking care of HIV-infected patients to validate risk assessment and prescriptions related to occupational or non-occupational HIV exposures. According to recommendations, 571 ‘PEP reference physicians’ were eligible in the 1999 survey.
Non-occupational exposure was defined as unprotected sex, sharing of injection equipment or any other exposure to potentially contaminated blood through needle injuries, bites or fights. PEP, based on a three-drug regimen, is offered following anal or vaginal intercourse with a person known to be HIV-positive or at increased risk of infection (e.g. injecting drug users, men who have sex with men) and following sharing of IDU equipment. PEP is not offered for oral sex with a partner of unknown serostatus. PEP can be offered, according to risk factors of HIV transmission (high infectivity of the source partner, sexually transmitted diseases, traumatic sexual intercourse or menstruation) following vaginal intercourse with a partner of unknown HIV status and for any oral sex with a partner known to be HIV-positive or at increased risk.
A questionnaire concerning the number and type of requests and prescriptions for non-occupational PEP during the previous 12 months and the characteristics of the last treated exposure was mailed to eligible physicians in July 1997 and again in July 1999. In 1999, information on severe antiretroviral drug side-effects (i.e., requiring hospitalization or life-threatening) including their nature, timing of occurrence and the type of drug regimen, was also recorded. To assess their prescription attitudes, physicians were presented with five hypothetical exposure scenarios and asked whether or not they would prescribe PEP, and why.
Proportions were compared by chi-squared tests. Probability of side-effects and 95% confidence intervals (CI) were obtained by the exact binomial method (Epiinfo, epitable software ).
Replies were received from 43% (678/1604) of the physicians in 1997 and from 54% (309/571) in 1999. The distribution of medical specialities of the respondents differed between the two studies, internal medicine physicians being less represented and infectious disease department physicians more represented in 1999 (reference physicians only) than in 1997 (respectively 39% versus 65% and 27% versus 22%;P < 10−3). However, the distribution of specialities was similar for respondents and non-respondents, both in 1997 and 1999. The vast majority of 1999 respondents (290/309; 94%) were aware of the 1998 recommendations.
Number and type of requests and prescriptions
A total of 390 requests and 165 prescriptions were reported in 1997 by the 678 responding physicians, of whom 74% had received no requests and 14% one request only during the previous 12 months. In 1999, the 309 reference physicians who responded reported 3187 requests and 1835 prescriptions; 18% had received no requests during the previous 12 months, 12% had one and 70% between two and 100.
The proportion of reports related to sexual exposure increased between 1997 and 1999, from 71% to 86% for requests (P < 10−6) and from 63% to 86% for prescriptions (P < 10−6). In 1999, the proportions related to sharing injection equipment (2% of requests and 3% of prescriptions) and to other contacts with potentially contaminated blood (11% of both requests and prescriptions) were lower than in 1997 (P < 10−6). The overall prescription rate increased from 42% (165/390) in 1997 to 58% (1835/3187) in 1999 (P < 10−6), with prescription following sexual exposure presenting the largest increase (from 37% to 58%;P < 10−6).
Description of the last HIV exposure treated
In 1997, 120 physicians provided details of their latest treated non-occupational HIV exposure, of which 68% were sexual exposures, 31% percutaneous exposures and 1% exposures by sharing injection equipment. In 1999, 226 treated non-occupational HIV exposures were described: 84% were sexual exposures, 14% percutaneous exposures (of which 24 were injuries with discarded syringe, bites, fights) and 1% exposures by sharing injection equipment. These distributions did not differ from that of all prescriptions reported either in 1997 or in 1999 (P = 0.54;P = 0.21).
Age and sex
Age and sex were not systematically reported in 1997. In 1999, 57% of latest treated patients were men and 68% were aged between 20 and 39 years. Among children (younger than 15 years), 11 were exposed through needle stick accidents (representing 50% of this type of exposure) and two were sexual assault victims. Among adolescents (15–17 years), 11 sexual exposures were reported, of which seven were assaults.
Among sexual exposures for which prophylaxis was prescribed in 1999, unprotected intercourse was more common (60%) than condom failure (40%). The opposite was observed in 1997 (respectively 30% and 70%;P < 10−5) (Table 1). Source serostatus was less likely to be known in 1999 (40% with known positive serostatus) than in 1997 (78%;P < 10−6). Knowledge of source serostatus was greater for exposures related to condom failure than for those related to unprotected sex, both in 1999 (55% versus 31%;P = 0.001) and in 1997 (91% versus 48%;P < 10−4) and did not differ according to sexual practices (P ≤ 0.04) (Table 1). After excluding sexual assault, vaginal intercourse accounted for almost two-thirds (62%) and anal intercourse for one-third of all exposures in both 1997 and 1999.
In 1999, 32 sexual assaults were described, in which the victims were 31 women and one man. Rape accounted for an increasing proportion of PEP after sexual exposure [7% (6/82) in 1997, 17% (32/191) in 1999;P = 0.04] (Table 1).
The proportion of non-sexual exposures treated decreased from 32% (38/120) in 1997 to 15% (35/226) in 1999 (P < 0.001). The source serostatus was known in 38% (14/38) of cases in 1997 and 14% (5/35) in 1999 (P = 0.05).
The latest prescription
Among the 224 treated patients in 1999, 92% received treatment within 48 h of the exposure, versus 87% in 1997 (P = 0.22;Table 2). There was a slight decrease in the proportion of patients treated more than 72 h after exposure: 2% in 1999 versus 9% in 1997 (P < 0.01). The delay did not differ according to sexual or non-sexual exposure either in 1997 or in 1999. However, for sexual exposures, the delay from exposure to prescription was shorter for condom failure than for unprotected sex both in 1997 [75% (43/57) treated < 24 h after condom failure versus 44% (11/25) for unprotected sex;P < 0.01] and in 1999 [83% (64/77) versus 57% (65/114), respectively;P < 0.001]. These delays were similar in 1997 and 1999 (for condom failure, P = 0.27; for unprotected sex, P = 0.24).
In 1999, combination therapies including a protease inhibitor (PI) or a non-nucleoside reverse transcriptase inhibitor (NNRTI) were the most common prescription (83% versus 46% in 1997;P < 10−6;Table 2), notably the combinations of zidovudine + lamivudine + indinavir (47%, 105/224) and zidovudine + lamivudine + nelfinavir (32%, 71/224). The dual combination zidovudine + lamivudine represented 14% (32/224) of prescriptions and no mono therapy was prescribed. Combination therapy with PI or NNRTI was more frequent in sexual than in other exposures both in 1997 [54% (44/81) versus 27% (10/38);P < 0.01] and in 1999 [87% (166/191) versus 60% (21/35);P < 10−4]. However this therapeutic choice was not influenced by the sex of the patient (in 1999), delay between exposure and treatment, type of sexual practice (anal or vaginal), whether intercourse was protected or not, knowledge of source HIV serostatus, or the physician's medical speciality. The preferred duration of treatment prescribed was 4 weeks particularly in 1999 (97% of cases versus 83% in 1997;P < 10−5;Table 2).
Severe side-effects of antiretroviral drugs
Among the 1835 prescriptions reported in 1999, 13 side-effects requiring hospitalization were described (Table 3), representing an estimated risk of 0.71% (95% CI, 0.38%–1.20%). Serious side-effects occurred only among patients under combination therapies including PI or NNRTI. The patients concerned were aged between 20 and 46 years and 77% were women. Nephrolithiasis was the most common event reported (six cases). Other severe events reported were three severe rashes, two toxic hepatitis, one cholecystitis and one haemolysis (Table 3). Seven side-effects occurred within 10 days of initiating therapy (range, 1–26 days). All side-effects were reversible after withdrawal of the drugs.
Side-effects occurred after indinavir treatment in 11 patients and after nevirapine in one patient; one patient received both indinavir and nevirapine (Table 3). Indinavir was included in 92% (12/13) of the combination therapies with side-effects reported compared with 44% (93/211) of combination therapies without side-effects reported (P < 10−3).
Prescription attitudes of physicians
To assess prescription attitudes, physicians were presented with five exposure scenarios (Table 4) and asked whether and why they thought there was or was not an indication for PEP.
Attitude by scenario
In both 1997 and 1999, a large majority of physicians would prescribe PEP in the case of a source known to be HIV infected (scenarios A and E). The rate of prescription for these scenarios increased significantly between 1997 and 1999 [from 93% to 99% in scenario A (P = 10−2) and from 89% to 99% in scenario E (P < 10−4);Table 4]. The main reason given was that the risk of transmission is high. The most important change between 1997 and 1999 was the increase of prescription in the case of a man having sex with a man of unknown HIV serostatus (scenario B), from 35% in 1997 to 82% in 1999 (P < 10−6;Table 4). In 1997, the main reasons for not prescribing in scenario B (among those having made comments, 155/404), were difficulty in risk assessment (34/155), continuous risky behaviour (18/155) and a threat for prevention (26/155). In 1999, the main reason for prescribing in this case (among those having made comments, 42/243), was the importance of taking into account the patient's perception of risk (12/42), provided that there was no continuous exposure (7/42).
In 1999, nearly half of physicians would prescribe PEP in a case of vaginal intercourse with a partner of unknown HIV serostatus (scenario C) compared with only 12% in 1997 (P < 10−6). Among physicians in favour of PEP in case of vaginal intercourse in 1999, the reasons most commonly cited (among those having made comments, 37/138) were that prescription should be given at the request of the patient (16/37), or if there were risk factors for transmission (10/37). Among those not in favour of PEP prescription, evidence of risk factors for transmission would change their attitude (21/43). In 1997, the main reasons for not prescribing in case of vaginal intercourse (among those having made comments, 183/542) were low risk (42/183) or difficulty of risk assessment (32/183).
About 25% of physicians in 1999 would propose PEP to a woman having performed unprotected oral sex with ejaculation with a casual male sex partner of unknown HIV serostatus (scenario D), while only 8% would have done so in 1997 (P < 10−6;Table 4). For this scenario, in 1999, request of the patient was again the most important factor (15/29) for PEP prescription, while among those not in favour of PEP prescription, evidence of risk factor (20/44) or the patient's request (12/44) would change their attitude. In 1997, reasons for not prescribing in this case were still focused on difficulty of risk assessment (59/158) and low risk (21/158), while few clinicians (17/84) would prescribe at the request of the patient.
Overall attitudes concerning PEP
Overall attitudes toward non-occupational PEP can be described from the response profiles to all scenarios. Physicians having responded (by yes or no) to every scenario (84% of respondents in 1997, 77% in 1999) were classified in four main profiles (Table 5). These main profiles can be separated into two categories: wide prescription, based mainly on the request of the patient (11% of physicians in 1997, 41% in 1999), and prescription based mainly on risk assessment (64% of physicians in 1997, 33% in 1999).
Wide prescription is composed of two groups: physicians who would propose PEP in every scenario (6% in 1997 and 22% in 1999 ) and those who would propose PEP in every scenario except oral sex (4.9% in 1997 and 18.8% in 1999). In the first group, the decision to treat depends mainly on the patient's request according to physicians’ comments, risk assessment being secondary (Table 5), while in the second both the patient's request and risk assessment are taken into account.
Prescription based mainly on risk assessment consisted of two other groups of physicians: those who would propose PEP either when the source is known to be HIV positive or when the patient has high risk practices (unprotected anal sex) (17% in 1997 and 24% in 1999) and those who would propose PEP only when the source is known to be HIV positive [the most frequent attitude in 1997 (47.3%) and one of the less frequent in 1999 (9.5%)]. This attitude is based on risk assessment only, according to physicians’ comments, the risk being considered maximum when the source is known to be HIV positive. Patient request doesn't play a major role. Physicians’ comments were essentially related to the level of risk.
Finally, in 1999 not a single physician would refuse to propose PEP in any of the scenarios, whereas 1.6% would have refused in 1997.
The physician panel of 1999 was restricted compared to that of 1997, but was nevertheless defined in the same way – physicians in a position to take care of persons exposed to HIV. Before the implementation of the 1998 recommendations, all physicians involved in the care of infected persons would prescribe PEP if they thought it was necessary. Since 1998, the recommendations require that a reference physician confirms and follows any prescription made after non-occupational HIV exposure. Reference physicians having participated in the 1997 survey didn't differ from other physicians in terms of attitudes (data not shown). The differences in prescriptions and attitudes observed between the two surveys cannot therefore be attributed to the modification of the study population but rather to the impact of the recommendations.
The retrospective nature of the study, which that could have resulted in less precision in the replies of requests and prescriptions reported by some physicians, and the low response rate of the two studies (respectively 43% and 54%) did not allow these numbers to be extrapolated to the whole country. Nevertheless, the comparison of the reported numbers between the two study periods suggests an increase in numbers of both of requests and prescriptions.
Whatever the real number of requests, it is clearly low with respect to the likely number of exposures per year. Therefore the French PEP programme probably reached only a small portion of the target population, as did other PEP programmes whether they were city-wide programmes run by a Department of Health as in San Francisco , very local programmes in big cities (New York City , Boston ), nationwide programmes (Switzerland ) or regional programmes (New South Wales, Australia ). In PEP programmes operated for at least 12 months, 401 individuals were treated during a 16-month period in San Francisco  and 110 in 19 months in Switzerland . However, a PEP programme should be only a small part of a comprehensive HIV prevention programme . Low access is not in itself a source of concern – rather it is the ability of the programme to reach the appropriate population that matters.
With only 79 (2%) requests reported in 1999, the French programme had little success in reaching the estimated 70 000 non-infected IDU, of whom 15–20% still share injecting material . Possible explanations for this finding include lack of awareness of the PEP programme among IDU, reluctance of health care providers in emergency rooms towards IDU, high frequency of exposures (for which a 28-day treatment makes little or no sense) and, in particular, a low priority among IDU taking into account their many other concerns (searching for drugs, injection material, food, accommodation). Results were similar in other programmes [12–16] that targeted IDU; PEP demand following sharing material represented between 2 and 4% of all demands, even in New York City (2%) . In the French programme, sexual exposure represented the main reason for PEP request, and increasingly so between 1997 and 1999 (86% of requests in 1999). The prescription rate for these exposures also increased (from 37% in 1997 to 58% in 1999). In other programmes, sexual exposures represented 85–96% of requests treated [12–14,16]. However, in several programmes, as in France, 7–15% of requests treated were very low risk exposures (no contamination proved) such as injury with discarded needles [14–16].
The analysis of last treated cases shows that the frequency of long delays (72 h and more) from exposure to treatment decreased in 1999, probably because the 1998 recommendations advised not to treat after 48 h. Delays observed in France seem to be shorter than those of the San Francisco programme (69% treated within 24 h in France, compared to a mean delay of 33 h in San Francisco ) where, as in other US programmes, patients are eligible for PEP up to 72 h after exposure. Among sexual exposures, the increase in the proportion of unprotected intercourse in 1999 and the corresponding decrease in that of condom failures could be interpreted as an effect of the 1998 recommendations. Indeed the recommendations indicated that the goal was to treat accidents or failures of prevention, thus allowing patients to speak more easily of risky behaviour and reluctant physicians to provide PEP to individuals whose sexual activities exposed them to HIV infection. The decreasing proportion of cases in which the source was known to be positive (from 78% in 1997 to 40% in 1999) could also result from the impact of the recommendations which allowed PEP prescriptions when the source status is unknown (but with factors increasing the risk of transmission). However, in the programmes implemented elsewhere, the proportions of such cases ranged from 33 to 47%, similar to the 40% we observed in 1999 [12–16]. Additional data from the French prospective study on PEP, implemented in July 1999, indicate that this proportion decreased further to 28% among 982 sexual exposures treated between July 1999 and May 2000 . Moreover, this proportion does not seem to be influenced by eligibility criteria: in Switzerland , where emphasis was placed on exposures to persons known to be positive, only 33% of sources were in fact known to be positive, whereas in Boston , where PEP eligibility included exposures to ejaculate, blood or vaginal secretions from an HIV positive or unknown status partner, a source known to be positive was reported in 37% of exposures.
Severe side-effects were reported exclusively among individuals receiving three drug combinations of two nucleoside analogues plus indinavir (12 cases) or nevirapine (two cases). The risk of 0.71% of severe side-effects is high for a 4-week treatment among healthy individuals and even if this risk is overestimated (physicians having observed side-effects might have better participated in the study), it is probably higher than the estimated risk of HIV transmission after a single act of vaginal intercourse with an HIV-positive partner, around 0.1% . Of note, our results are consistent with the nine severe adverse symptoms observed in the French prospective PEP study  among 982 PEP recipients between July 1999 and May 2000 (0.92%; 95% CI, 0.42–1.73). In our study the link between specific drugs and side-effects is based on clinicians’ reports: we did not perform a detailed review of all patient medical records. However, the fact that all reported side-effects were within the toxicity profile of the prescribed drugs and occurred among previously healthy individuals with a temporal sequence after drug prescription, pleads in favour of a link. Even if they needed hospitalization, side-effects reported in our study were not life-threatening. However, life-threatening side-effects have been reported among prophylaxis recipients with nevirapine in the USA  and with indinavir in Switzerland .
The French recommendations did not include standardized treatment algorithms based on severity of exposure, but simply indicated: ‘‘Medication regimen choices must be made case by case …, according to the recommendations of the French expert group on use of antiretroviral drugs’'. This latter group proposed a highly active regimen of two nucleoside analogues plus a PI, as first line treatment: ‘‘this is the most powerful combination which should be considered first’', even though, in the vast majority of cases the HIV status of the source was unknown . This may result from ambiguity as to the aims of PEP treatment, namely prevention of infection by treatment of exposed persons as compared to early treatment of infection with highly active regimens. This ambiguity was also apparent in the 1998 CDC guidelines for occupational exposures to HIV  which, concerning the timing of PEP initiation stated: ‘‘Initiating therapy after a long interval may be considered, … even if infection is not prevented, early treatment of acute infection may be beneficial’'. However, the very recent revision (June 2001 ) no longer refers to treatment, the second part of the sentence having been replaced by ‘‘for exposures that represent an increased risk for transmission’'. Because of the potential toxicity of PEP, the 1998 (and 2001) CDC guidelines stated ‘‘there is insufficient evidence to recommend a highly active regimen for all HIV exposures’’ and two regimens, ‘basic’ (two drugs) and ‘expanded’ (three drugs) for exposures that carry an increased risk for transmission, are recommended. Guidelines from the San Francisco non-occupational PEP programme clearly aimed at prevention of infection . This is reflected in the medication algorithm that recommends the addition of a PI only if the source reports detectable viral load while on therapy and subject to considerations of added potency versus added toxicity . Consequently, two-drug combinations represent 86% of PEP prescriptions in San Francisco , whereas in other programmes [13–16], three-drug combinations account for 74–85% of prescriptions. Thus the use of three-drug combinations with PI or NNRTI as first-line therapy should be reconsidered taking into account toxicity and the aim of prevention of infection rather than treatment [26,27]. Nucleoside analogues should be preferable according to their pharmacokinetic properties to prevent infection , whereas a three-drug combination should be more secure, in certain circumstances, according to antiretroviral resistance profiles among sources in a country .
Our analysis of prescription attitudes indicates that the recommendations have had an impact on physicians’ behaviour. The most frequent attitude in 1999, wide prescription based mainly on patient request (41% of physicians, compared with only 11% in 1997), fits with the 1998 recommendations, which note in preamble ‘‘It is important to consider the perception of risk leading a person to request prophylaxis’’ . However, the recommendations include also a wide discussion of risk assessment and do not advise treatment for low risk practices when the status of the source is not known (other than sexual assault), unless there are risk factors (e.g., sexually transmitted diseases). Thus the high frequency of prescription, regardless of whether or not the status of the source is known, might indicate that the concern of the government for equal access to this treatment for everyone (expressed by the importance given to the request of the person) had been more considered than the necessity of risk assessment. Another interpretation could also be proposed connected to the French context of physicians’ concerns about their own liability in the post-blood scandal era. In 1999, no clinician would refuse to provide PEP, whatever the exposure. In 1997, on the other hand, the most frequent clinician attitude (64% of clinicians) was to treat only those whose source was known to be HIV positive or to have high risk behaviour, an attitude much less frequent in 1999 (33%). Thus the very existence of official recommendations, independently of their content, may have induced clinicians to treat more than they otherwise would. In the USA, the ‘considerations’ for the use of antiretroviral drugs for non-occupational PEP are relatively opposed to the French recommendations with respect to the importance to be given to the request of the patient: ‘‘PEP should never be administered routinely or solely at the request of a patient’’ . Such a statement may help physicians to exercise their judgement with less fear of possible legal consequences and to resist pressure from patients to offer treatment in cases where it is not appropriate.
Despite the limits of this retrospective study, we conclude that recommendations introduced in 1998 probably improved clinician acceptability of non-occupational PEP in France and broadened indications for PEP by giving greater consideration to the request of the patient. This has resulted in more frequent treatment of less risky exposures. The high number of severe side-effects reported among healthy individuals during a 12-month period needs to be taken into account in a benefit-risk analysis of PEP, not thoroughly considered in the climate of therapeutic optimism prevailing when the recommendations were formulated. Used as a prevention tool, PEP requires an approach different to therapy from that used for the treatment of an established HIV infection. Taken together, these findings plead in favour of a revision of the current French recommendations, to ensure that for exposures treated the risk of HIV transmission outweighs the risk of severe side-effects.
1. Ministère de l'Emploi et de la Solidarite. Diagnostic précoce, traitement antirétroviral après exposition et dépistage de l'infection par le VIH. Paris: Ministère de l'Emploi et de la Solidarité; 1998: 39–54.
2. Conner EM, Sperling RS, Gelber R. et al. Reduction of maternal-infant transmission of human immunodeficiency virus type 1 with zidovudine treatment. N Engl J Med 1994, 331: 1173–1180.
3. Cardo DM, Culver DH, Ciesielski CA. et al. A case control study of HIV seroconversion in health care workers after percutaneous exposure. N Engl J Med 1997, 337: 1485–1490.
4. Tsai C-C, Follis KE, Sabo A. et al. Prevention of SIV infection in macaques by (R)-9-(2-phosphonylmethoxypropyl) adenine. Science 1995, 270: 1197–1199.
5. Bottiger D, Johansson N-G, Samuelsson B. et al. Prevention of simian immunodeficiency virus, SIVsm or HIV-2 infections in cynomolgus monkeys by pre- and postexposure administration of BEA-005. AIDS 1997, 11: 157–162.
6. Lamontagne F, Pillonel J, Laporte A. Prophylaxie après exposition non professionnelle au VIH : Expérience et attitude des médecins hospitaliers. Bull Epidémiol hebd 1998, 42: 183–185.
7. Katz MH, Gerberding JL. Postexposure treatment of people exposed to the human immunodeficiency virus through sexual contact or injection-drug use. N Engl J Med 1997, 336: 1097–1100.
8. Centers for Disease Control. Management of possible sexual, injecting-drug-use, or other nonoccupational exposure to HIV, including considerations related to antiretroviral therapy Public Health Service statement. MMWR 1998, 47 (RR–17): 1–14.
9. Office fédéral de la santé publique. Recommandations préliminaires pour la prophylaxie postexposition du VIH hors environnement médical. Bulletin 1997, 50: 4–7.
10. Ministère de l'emploi et de la solidarite. Circulaire DGS/DH/DRT/DSS n°98/228 du 9 avril 1998 relative aux recommandations de mise en œuvre d'un traitement antirétroviral après exposition au risque de transmission du VIH. Paris: Ministère de l'Emploi et de la Solidarité; 1998.
11. Dean AG, Dean JA, Coulombier D. et al. Epiinfo, version 6: a word processing, database and statistics program for public health on microcomputers. Atlanta: Centers for Disease Control and Prevention; 1995.
12. Khan JO, Martin JN, Roland ME. et al. Feasability of postexposure prophylaxis (PEP) against human immunodeficiency virus infection after sexual or injection drug use exposure: The San Francisco PEP study. J Infect Dis 2001, 183: 707–714.
13. Cadman J, Torres R, Georges MC, Bartell L, Corbett N. Challenges in designing the first nonoccupational post-exposure prophylaxis (PEP) pilot program and registry in NYC. XIII International Conference on AIDS. Durban, July 2000 [abstract 4309].
14. Kwong J, Mayer K, Peterson N, Appelbaum J, Lasalvia T, Boswell S. Non-occupational Post-Exposure Prophylaxis (NPEP) at a Boston Community Health Center. XIII International Conference on AIDS. Durban, July 2000 [abstract 4310].
15. Bernasconi E, Ruef C, Jost J, Francioli P, Sudre P. National registry for non-occupational post HIV exposure prophylaxis in Switzerland: two years results. XIII International Conference on AIDS. Durban, July 2000 [abstract 4312].
16. Correll P, Smith DE, Kippax S, Hendry O, Grulich AE. Non occupational HIV post exposure prophylaxis (PEP) in Australia. XIII International Conference on AIDS. Durban, July 2000 [abstract 4420].
17. Lurie P, Miller S, Hecht F, Chesney M, Lo B. Postexposure prophylaxis after nonoccupational HIV exposure, clinical, ethical and policy considerations. JAMA 1998, 280: 1769–1773.
18. Valenciano M, Emmanuelli J, Lert F. Unsafe injecting practices among attendees of syringe exchange programs in France. Addiction 2001, 96: 597–606.
19. Lot F, Larsen C, Baum-Parmentier V, Laporte A. Sexual HIV post-exposure prophylaxis (PEP) in France.Eighth Conference on Retroviruses and Opportunistic Infections., Chicago, February 2001 [abstract 226].
20. Leynaert B, Downs AM,, DE Vincenzi I for the European study group on heterosexual transmission of HIV. Heterosexual transmission of human immunodeficiency virus. Variability of infectivity throughout the course of infection. Am J Epidemiol 1998, 148: 88–96.
21. Centers for Disease Control and Prevention. Serious adverse events attributed to nevirapine regimens for postexposure prophylaxis after HIV exposures - Worldwide, 1997–2000. MMWR 2001, 49 (51 and 52): 1153–1156.
22. Ministère de l'Emploi et de la Solidarite. Stratégies d'Utilisation des Antirétroviraux dans l'infection par le VIH. Paris: Ministère de l'Emploi et de la Solidarité; 1998: 102–104.
23. Centers for Disease Control and Prevention. Public health Service Guidelines for the management of Health-care workers exposures to HIV and recommendations for postexposure prophylaxis. MMWR 1998, 47 (RR–7): 1–33.
24. Centers for Disease Control and Prevention. Updated Public Health Service guidelines for the management of occupational exposures to HBV, HCV, and HIV and recommendations for postexposure prophylaxis. MMWR 2001, 50 (RR–11): 1–53.
25. Roland M, Klausner JD, Bangsberg DR, Franses K, Stansell J, Katz M. Non-occupational post-exposure prevention (PEP): integrating research findings into two clinical settings to provide a comprehensive service. XIII International Conference on AIDS. Durban, July 2000 [abstract 4326].
26. Struble KA, Pratt RD, Gitterman SR. Toxicity of antiretroviral agents. Am J Med 1997, 102 (5B): 65–67.
27. Parkin JM, Murphy M, Anderson J, El-Gadi S, Forster G, Pinching AJ. Tolerability and side-effects of post-exposure prophylaxis for HIV infection. Lancet 2000, 355: 722–723.
28. Flexner CW. Principles of clinical pharmacology in postexposure prophylaxis. Am J Med 1997, 102 (5B): 32–38.
29. Prevot MH, Descamps D, Migueres B, et al. Antiretroviral resistance profiles in source patients during occupational exposure to blood in French hospitals. XIII International Conference on AIDS. Durban, July 2000 [abstract 4321].
© 2002 Lippincott Williams & Wilkins, Inc.