clinicoepidemiologic characteristics, prognostic ... - Semantic Scholar

Am. J. Trop. Med. Hyg., 58(4), 1998, pp. 436–443 Copyright q 1998 by The American Society of Tropical Medicine and Hygiene

CLINICOEPIDEMIOLOGIC CHARACTERISTICS, PROGNOSTIC FACTORS, AND SURVIVAL ANALYSIS OF PATIENTS COINFECTED WITH HUMAN IMMUNODEFICIENCY VIRUS AND LEISHMANIA IN AN AREA OF MADRID, SPAIN ROGELIO LOPEZ-VELEZ, JOSE A. PEREZ-MOLINA, ANTONIO GUERRERO, FERNANDO BAQUERO, JESUS VILLARRUBIA, LUIS ESCRIBANO, CARMEN BELLAS, FRANCISCO PEREZ-CORRAL, AND JORGE ALVAR Clinical Microbiology and Infectious Diseases Department, Hematology Department, Pathology Department, and Biostatistics Department, Hospital Ramon y Cajal, Madrid, Spain; Instituto Carlos III, Madrid, Spain

Abstract. From 1987 to 1995, a retrospective case study was conducted at the Ramon y Cajal Hospital in Madrid, Spain, a public teaching hospital with 1,100 beds, to determine the clinicoepidemiologic characteristics, survival, and prognostic factors of patients with visceral leishmaniasis (VL) and human immunodeficiency virus (HIV) infection. The prevalence of VL in HIV1 patients compared with HIV2 patients was studied. Epidemiologic, clinical, and parasitologic characteristics, as well as the effects of treatment, prognosis, and survival in 54 HIV1 patients (90 episodes) with VL were defined. Comparative survival studies among patients with and without acquired immunodeficiency syndrome (AIDS)–defining criteria and multivariate analysis of survival risk factors were performed. The prevalence of VL in patients with AIDS was much higher than in immunocompetent individuals. In spite of a good initial response to treatment for VL, 60.6% of the patients had relapsed by the end of one year. Mortality from the first episode was 18.5%, and 24% died in the first month after diagnosis of any VL episode. The mean survival of the 29 patients who died was 10.27 months. Survival in patients with and without AIDS at the time of the first episode of VL was compared at 30 months: 53.7% versus 20.5% (P 5 0.00149). We found no significant difference (P 5 0.24) in the survival of HIV1 patients who had died of VL without AIDS at the time of the first episode of VL compared with those of a control group of 413 dead patients with AIDS without VL. A diagnosis of AIDS at the time of the first episode of VL and thrombocytopenia were the only risk factors found related to survival. We conclude that in AIDS patients, VL is a recurrent disease that is highly prevalent and whose clinical course is modified by HIV. Leishmaniasis is a parasitic disease caused by infection with protozoa of the genus Leishmania. It is transmitted by the bite of infected phlebotomine sand flies and is distributed throughout the New and the Old World. Leishmania infantum is responsible for visceral leishmaniasis (VL) in the Mediterranean basin (where this disease is endemic). Since the mid 1980s, the number of cases of VL in patients infected with human immunodeficiency virus (HIV) has increased.1–3 From a global epidemiologic viewpoint, two parallel trends are alarming: the ruralization and spread of the HIV pandemic4 and the urbanization and dissemination of leishmaniasis.5 Current population movements, either for socioeconomic, tourism, or military reasons, result in the exposure of more nonimmune people to the zoonotic cycle of Leishmania, which could lead in the near future to an increase in the cases of HIV and Leishmania coinfection, even in areas nonendemic for VL. In southern Europe, between 25% and 75% of cases of adult VL occur in HIV1 patients and 2–9% of patients with acquired immunodeficiency syndrome (AIDS) have VL.6 In HIV1 patients, VL shares certain characteristics with other opportunistic diseases: a high prevalence, an association with severe immunosuppression, a frequent oligosymptomatic or atypical manifestations, often not detected by routine diagnostic tests, and a poor response to treatment, with frequent relapses.7, 8 The objectives of this study were to determine the clinicoepidemiologic characteristics, survival, and prognostic factors of patients coinfected with HIV and Leishmania.

hospital with 1,100 beds serving a well-defined area in northern Madrid (officially called area 4). This hospital is the only source of tertiary level care for the population of 550,000 living in that area. From March 1987 through May 1995, patients diagnosed with VL were reviewed. An active search for Leishmania was conducted in all patients with prolonged fever and/or hepatosplenomegaly and/or hematologic cytopenias. Data were collected on demographic information, stage of HIV infection, diagnosis of VL, other previous or concomitant conditions, clinical presentation, laboratory test results, parasitologic data, treatment received, secondary effects, clinical course, and survival. The survival of patients with and without AIDS-defining criteria at the time of the first episode of VL was compared. Survival of patients who had died of VL without AIDS at the time of the first episode of VL were compared with those of a randomly selected control group of 413 dead patients diagnosed with AIDS within the same time period and from the same geographic area but without VL. Ethical considerations and the study design were reviewed and approved by Hospital Institutional Review Board. All patients provided informed consent for invasive diagnostic procedures as well as for antiparasitic treatment. Case definition and diagnosis of VL. Infection with HIV21 and the stage and diagnosis of AIDS were defined according to the Centers for Disease Control (Atlanta, GA) criteria of 1992.9 Pulmonary tuberculosis, repeated pneumonia, and invasive carcinoma of the uterine cervix have been officially included as confirmatory of AIDS in Spain since January 1994. Stages A3 and B3 are not classified as confirmatory of AIDS in Europe10 (clinical category A is defined as asymptomatic HIV infection, persistent generalized lymphadenopathy, or acute primary HIV illness; cate-

PATIENTS AND METHODS

Patients. This retrospective study was performed at the Ramon y Cajal Hospital in Madrid, Spain. This is a public

436

437

LEISHMANIASIS AND HIV COINFECTION

gory B is defined as symptomatic non-A, non-C conditions, such as bacillary angiomatosis, persistent vulvovaginal candidiasis, oropharyngeal candidiasis, severe cervical dysplasia, carcinona in situ, or constitutional symptoms; category C is defined as the occurrence of opportunistic infections or neoplasms; and CD4 range 3 is defined as a CD4 lymphocyte count , 200/mm3). A diagnosis of VL was made if Leishmania amastigotes were observed or promastigotes were isolated in parasite cultures (NNN medium) from biopsy-aspirate or buffy coat samples. Serology was considered positive if titers $ 1:80 measured by a commercial indirect immunofluorescence antibody test (IFAT) (Bios, Munich, Germany). Leishmania isolates were characterized by isoenzyme typing by starch gel electrophoresis.11, 12 The electrophoretic patterns were compared with those from Leishmania species markers and intra-L. infantum species reference stocks (MHOM/FR/78/LLM-75). Treatment, response criteria, and adverse effects. The following three treatment regimens were used. The first was a pentavalent antimonial (meglumine antimoniate) at a dose of 20 mg/kg/day given intravenously (iv) or intramuscularly for 28 days. These were combined in three cases with allopurinol, 900 mg/day orally, or recombinant human interferon-gamma (rHIFNg), 100 mg/m2 subcutaneously for 28 days (RU 42369; Roussel-UCLAF, Romainville, France). The second was amphotericin-B at total doses of at least 15 mg/kg iv. The liposome formulation (Ambisomet; Nexstar, San Diego, CA) was given to four patients. The third was pentamicline isethionate at doses of 4 mg/kg/day iv for 28 days. The clinical response was defined as good if after a complete course of treatment the patient was afebrile with improvement of the pancytopenia and hepatosplenomegaly, and was asymptomatic three weeks after the end of treatment; poor if the fever, cytopenia, or hepatosplenomegaly persisted after 21 days of treatment; and not assessable if the patient did not complete treatment or died before completing treatment. A parasitologic control study (staining and parasitic culture of bone marrow) was performed (without any selective criteria) on some patients one month after completing the treatment. A parasitologic cure was indicated if Leishmania was not identified after staining and culture of a bone marrow aspirate. Evaluation of adverse reactions to antileishmanial drugs was carefully carried out by two independent observers every five days, given that some patients were seriously ill and were receiving other therapeutical agents. Side effects were monitored using the following tests: complete blood count, coagulation profile, electrolytes, renal, hepatic, and pancreatic function, urine sediment, and electrocardiogram (ECG). The following were considered serious adverse effects: anemia (defined as a 25% reduction in the hematocrit), renal toxicity (a three-fold increase in the normal level of serum creatinine), hepatic toxicity (a 10-fold increase in the base values of the transaminases), and hyperamylasemia (a twofold increase in normal serum amylase values). Statistical analysis. The nonparametric Wilcoxon twotailed test was used to compare the mean of two samples without normal distribution. Percentages were compared using the exact binomial method. Survival curves were estimated using the Kaplan-Meier method. Differences between

survival curves were evaluated using the log-rank test. Survival at 30 months was estimated: confidence intervals (CIs) were calculated using logarithmic transformation. To identify independently significant factors that could influence survival or relapses, a Cox stepwise proportional hazard model was applied in analysis of the 54 patients. The maximum model included seven variables measured at the first episode of VL: sex, presence of AIDS-defining illness, CD41 lymphocyte count, hemoglobin level, neutrophil count, platelet count, and total serum proteins. A backward procedure strategy model was used to obtain the best models for the association of prognostic factors with survival. The fit of the models was assessed by the log-likelihood ratio test. All P values were two-sided and values 0.05 or less were considered statistically significant. The PRESTA software package was used for the database and all statistical procedures.13 RESULTS

Epidemiologic features, HIV clinical stage, and immunologic status. During the 98-month study period, 66 newly diagnosed cases of VL were observed in the population served by the hospital. Ten (15.2%) were not infected with HIV (seven immunocompetent and three transplant patients) and 56 (84.2%) were HIV1. For each new case of VL diagnosed in an HIV2 patient, 5.6 new cases were diagnosed in HIV1 patients. Ninety-two separate episodes of VL were recorded in the 56 HIV1 patients during the study period. Of these, 90 episodes (54 first episodes and 36 relapses) and 54 patients were assessed in this study. At the time of the first episode of VL, 25 (42.3%) of 54 patients met the European diagnostic criteria for AIDS, but most patients were strongly immunocompromised. Age, sex, risk factors for HIV infection and immunologic state are shown in Table 1. Seven cases of VL were diagnosed in immunocompetent patients; the prevalence of VL in this group of patients was 1.2 per 100,000 (7/550,000). Twenty-five cases of VL were diagnosed among the HIV1 patients meeting AIDS criteria. At the end of the study period, 898 AIDS cases from the same area had been diagnosed at our hospital, and the prevalence of VL in this group was 2,784 per 100,000 (25 of 898). These data suggest that in this area of Madrid, VL was much more common in AIDS patients than in immunocompetent ones. Clinical features and laboratory data. The clinical characteristics of the episodes are shown in Table 2. A comparison of the laboratory results of the first episode and the first relapse is shown in Table 3. Diagnostic procedures. A study of bone marrow was the procedure most often used, being performed in 73 episodes. Isolation of the parasite in the buffy coat blood culture was attempted in 19 episodes. A spleen aspirate was carried out in 11 episodes: in nine patients the previous bone marrow aspirate was negative and the remaining two patients refused a bone marrow study. A liver biopsy was performed in five patients and gave an indication of a previous negative bone marrow study result in only one case (one patient had already been diagnosed by a bone marrow study and the biopsy was taken post-mortem in the other three). Serology

438

LOPEZ-VELEZ AND OTHERS

TABLE 1 Epidemiologic features, human immunodeficiency virus (HIV) clinical stage, and immunologic status in 54 HIV-positive patients at the first episode of visceral leishmaniasis Age, years (mean 6 SD) Sex Male Female Risk factor for HIV-1 infection Intravenous drug abuser Homo-bisexual contact Heterosexual contact Blood transfusion Unknown HIV infection stage* A B C CD41 lymphocyte count 3 106/L† . 500 201–500 101–200 50–100 , 50

31 6 7.5 46 (85.1%) 8 (14.9%) 43 4 3 2 2

(79.6%) (7.4%) (5.6%) (3.7%) (3.7%)

16 (29.6%) 13 (24.1%) 25 (46.3%) 1 4 9 15 18

(2.1%) (8.6%) (19.1%) (31.9%) (38.3%)

* Clinical category A is defined as asymptomatic HIV infection, persistent generalized lymphadenopathy, or acute primary HIV illness. Category B is defined as symptomatic, non-A, non-B conditions such as bacillary angiomatosis, persistent vulvovaginal candidiasis, oropharyngeal candidiasis, severe cervical dysplasia, carcinoma in situ, or constitutional symptoms. Category C is defined as esophageal/tracheal/bronchial candidiasis, extrapulmonary coccidioidomycosis, invasive cervical cancer/ chronic intestinal cryptosporidiosis/cytomegalovirus retinitis/disseminated disease, HIV encephalopathy, chronic herpes simplex virus infection, mucocutaneous ulcer/bronchitis, pneumonia, chronic disseminated extrapulmonary histoplasmosis, chronic isosporiasis, Kaposi’s sarcoma, Burkitt’s lymphoma, extrapulmonary Mycobacterium avium/M. kansasi, pulmonary/extrapulmonary tuberculosis, Pneumocystis carinii pneumonia, recurrent pneumonia, progressive multifocal leukoencephalopathy, recurrent Salmonella bacteremia, cerebral toxoplasmosis, or wasting syndrome due to HIV. † The CD41 lymphocyte count was available for 47 (87.043%) patients. ,200 5 42 (89.3%).

was performed in 73 episodes. A rectal and a duodenal biopsy were performed on one of the patients with chronic diarrhea in different episodes and Leishmania amastigotes were observed in both samples. Of the 90 episodes of VL, a diagnosis was made after a bone marrow study in 69 episodes, after a spleen puncture in nine episodes, after staining and/or culture of the buffy coat in eight episodes, after autopsy in three episodes, and after liver biopsy in one episode. The sensitivities of the diagnostic techniques used are shown in Table 4. There was no significant difference in the results of the diagnostic tests between the first episodes and the relapses. Twelve L. infantum isolates were characterized. Only four of 16 enzyme loci studied showed polymorphism. Zymodeme MON-1 was identified in three stocks, MON-24 in five, MON-29 in one, MON-34 in two, and MON-183 in one. Response to treatment. Response to treatment was evaluated in 71 of 90 episodes (78.9%). The remaining 19 episodes were not assessible for results because 10 patients died during treatment, three were never treated because they were diagnosed at autopsy, and six were treated with a suboptimal total dose of antiparasitic drugs. The response to each antiparasitic agent was as follows. A pentavalent antimonial was administered in 51 (71.9%) of 71 episodes. Allopurinol was also given in three cases and rHIFNg in two. The response was good in 42 (82.4%) and poor in nine (17.6%). Amphotericin B was administered in 17 (23.9%) of 71 episodes.

TABLE 2 Episodes of visceral leishmaniasis and human immunodeficiency virus (HIV) coinfection: clinical features and concomitant opportunistic infections Characteristic

Clinical presentation Visceral leishmaniasis Visceral and cutaneous leishmaniasis* Clinical findings Hepatomegaly Fever Splenomegaly Hepatosplenomegaly Constitutional syndrome† Lymphadenopathy Dry cough Diarrhea Hemorrhagic diathesis‡ Concomitant opportunistic infections§ Disseminated cytomegalovirus Central nervous system toxoplasmosis Esophageal candidiasis Pneumocystis carinii pneumonia Extrapulmonary tuberculosis HIV-related encephalopathy Disseminated Mycobacterium avium complex

No./no. available (%)

87/90 (96.7) 3/90 (3.3) 68/84 72/90 59/84 57/84 28/90 10/84 9/90 7/90 6/90

(80.9) (80) (70.2) (67.8) (31.1) (11.9) (10) (7.7) (6.6)

4/90 3/90 3/90 2/90 1/90 1/90 1/90

(4.4) (3.3) (3.3) (2.2) (1.1) (1.1) (1.1)

* Presented as a single cutaneous lesion (oriental sore). † Asthenia, fatigue, and loss of .10% of the body weight in the last month. ‡ Laboratory evidence of disseminated intravascular coagulation or major hemorrhagic manifestations. § Newly diagnosed at the same time as visceral leishmaniasis.

The response was good in 14 (82.3%) and poor in three (17.7%). Pentamidine was administered in three (4.2%) of 71 episodes. The response was good in two (66.6%) and poor in one (33.3%). A parasitologic control study (staining with Giemsa and culture) of the bone marrow was carried on 16 (22.5%) of 71 of the therapeutic regimens. Fourteen had been treated with pentavalent antimony and two with amphotericin B. The result was positive in three (18.7%) patients from the group treated with pentavalent antimony (combined with rHIFNg in one). The relapse rate at six months was 27% (10 of 37) and by one year was 60.6% (20 of 33). The mean period before occurrence of the first relapse was 9.3 months (95% CI 5 7.5–11.1) and from the first to the second relapse was 4.7 months (95% CI 5 3.7–6.1) (P 5 0.035). Six patients received secondary prophylaxis with pentamidine isethionate after the second episode. This was administered at a dose of 4 mg/kg iv at two weeks in three patients and four weeks in three others. There were no relapses during the follow-up (average 5 8 months, range 5 3–12 months). One of the patients stopped prophylaxis after 11 months and relapsed seven weeks later. Drug side effects appeared in 33 (34.6%) of the 95 therapeutic courses assessed (71 complete regimens and 24 incomplete). The incidence of serious reactions was as follows. In those receiving pentavalent antimonials (21 [30.8%] of 63 cases), hyperamylasemia was observed in eight (11.7%) patients, whether or not this was accompanied by symptoms of pancreatitis, changes on the ECG (increased Q-Tc and A-V blocks) were seen in eight (11.7%) and ar-

439


TABLE 3 Laboratory findings at first episode and first relapse of visceral leishmaniasis in human immunodeficiency virus–coinfected patients First episode Mean (no.)

Analytical data

Hemoglobin, g/dL Leukocytes/mm

Neutrophils/mm3 Platelets 3 10 /mm 3

Range

9.6 (50) 2,114 (49) 1,275 (47) 120.8 (50) 7.3 (37) 2.74 (21) 96.3 (47)

3

3

Serum proteins, g/dL IgG, G/dL CD4 lymphocytes/mm3

First relapse

4–13.7 160–4,200 160–4,200 30–298 4.7–11.2 0.94–9.02 3–702

Mean (no.)

10.2 (20) 2,388 (19) 1,617 (17) 107.3 (19) 7.6 (13) 4.35 (5) 41.0 (15)

Range

P*

5.1–13

NS

500–7,200

NS

200–5,900

NS

21–183

NS

6.5–13.2

NS

1.27–12.70

NS

1–180

0.029

* NS 5 not significant.

rhythmia in torsade de pointes with cardiac arrest and death was seen in one. In those receiving amphotericin B (nine [42.8%] of 21 cases), anemia was observed in six (28.5%) patients, with four of them requiring transfusion, and renal failure was seen in five (23.8%), with one of them developing diabetes insipidus. It is worth noting that patients did not develop either anemia or renal insufficiency on any of the four occasions in which liposomal amphotericin B was used. In those receiving pentamidine (three [50%] of six cases), insulin-dependent diabetes was observed in one patient and renal insufficiency in was seen in two (associated with leukopenia in one case). Survival. Death occurred during treatment of the first episode in 10 (18.5%) of 54 patients (average time to death 5 16 days) and 13 (24%) had died in the first month after diagnosis. Mean survival of the 29 patients who died was 10.27 months (95% CI 5 8.23–12.31). Long-term survival depended mainly on whether or not TABLE 4 Comparative parasitologic diagnostic methods in first episode and relapses

Diagnostic method

Serology Blood culture Liver biopsy Bone marrow aspirate Bone marrow biopsy Bone marrow culture Spleen aspirate Spleen culture

First episode

Relapses

Total

No. positive/ no. available (%)

No. positive no. available (%)

No. positive/ no. available (%)

15/47 (31.9) 3/9 (33.3) 3/4 (75) 28/45 (62.2) 24/35 (68.5) 29/34 (85.2) 5/5 (100) 4/4 (100)

12/26 (46.1) 5/10 (50) 0/1 (0) 18/28 (64.2) 16/23 (69.5) 20/26 (76.9) 6/6 (100) 5/5 (100)

27/73 (36.9) 8/19 (42.1) 3/5 (60) 46/73 (63) 40/58 (68.9) 49/60 (81.6) 11/11 (100) 9/9 (100)

the patient met the criteria for AIDS when VL was first diagnosed (Figure 1): at 30 months, survival was 53.7% (95% CI 5 26.4–74.8) in the group of 28 patients with VL without AIDS and 20.5% (95% CI 5 6.6–39.6) in the 26 patients with VL with AIDS (P 5 0.00149). We found no significant difference in the survival of patients dying of VL without AIDS criteria on the first episode (P 5 0.24) compared with that of the control group. The multivariate analysis showed that the only variables that had a significant effect on patient survival were the diagnosis of AIDS at the time of the first episode of VL (RR 2.42, 95% CI 5 1.040–5.650, P 5 0.0378) and a low platelet count (risk ratio [RR] 5 1.006; 95% CI 5 1.000–1.013, P 5 0.0434), with each reduction of 10,000 platelets increasing the mortality risk ratio by 6%. DISCUSSION

Until 1985, in Spain most patients with VL were immunocompetent children.14 Since then, nearly 80% of the cases have been in immunocompromised adult patients. Whereas leishmaniasis affects 0.3 of every 100,000 inhabitants in the total Spanish population, this increases to 2–6% in Spanish AIDS patients.15 A similar situation can be observed in southern Europe, where 2–9% of the AIDS patients develop VL.6 In this study, the prevalence of VL in AIDS patients was 2,320 times greater than in immunocompetent patients, showing a strong association between VL and HIV infection. We believe that there could be two possible reasons for this association. The first is increased susceptibility. Clinical VL represents the tip of the iceberg of the Leishmania-infected immunocompetent population, where only 1–6.5% of the infected patients develop clinical VL.5 In hyperendemic foci, up to 38% of the population have been in contact with Leishmania.16 Therefore, there are many cases of asymptomatically infected individuals who in the event of their becoming immunocompromised, would manifest the disease as a result of reactivation of a latent infection. Thus, the spread of HIV could indirectly reflect the latent prevalence of leishmaniasis in a population. Leishmaniasis in HIV1 patients

440


FIGURE 1. Long-term survival of 54 human immunodeficiency virus–positive patients with visceral leishmaniasis (VL) with (26 patients) or without (28 patients) acquired immunodeficiency syndrome (AIDS)–defining criteria on diagnosis after the first episode of VL.

could also be the result of a higher rate of clinical disease immediately following primary infection, given that coinfected patients may lack a protective immunologic defense. Leishmania has been recovered from the peripheral blood of HIV1 patients with VL,17, 18 and the natural vector19 is susceptible to infection by the parasite when it is feed with blood from coinfected patients in the laboratory. The quantity of blood that can be shared in syringes by drug users is greater than that sucked by the vector when biting (0.5 ml). It appears possible that in the population group, VL would be more prevalent than in other HIV groups, and the zoonotic cycle could be replaced by an anthroponotic one. In fact, in this series of patients 79.6% (95% CI 5 67.1–88.2) were intravenous drug users. This proportion is higher than the 68.8% of AIDS cases recorded in Madrid that belong to the intravenous drug users risk group.20 Although this difference is not statistically significant, it could suggest that the exchange of needles could aid transmission of the parasite.21 In Spain, L. infantum is responsible for both cutaneous and visceral leishmaniasis. In dogs and in most immunocompetent cases of VL, the L. infantum zymodeme isolated is MON-1, whereas in HIV1 populations, there is considerable variability in the zymodemes involved.22, 23 Our study revealed not only viscerotropic but also dermatotropic stocks, although none of these were isolated from the skin. In three episodes (typical Oriental sore on the face of one patient, and on the limbs of two patients) a cutaneous infection predated the visceral condition, which could be considered as visceralization secondary to an initial cutaneous lesion. In accordance with other studies,2, 3, 8 the clinical presentation in coinfected patients does not differ greatly from that of immunocompetent patients (fever and/or pancytopenia and/or hepatosplenomegaly). Because of the severely immunocompromised condition of these patients, the parasite can spread to tissues that it usually does not inhabit due to the absence of an efficient immune response system24 and colonize in atypical locations.7 Ten percent of our HIV1 patients presented with a

dry cough and 7.7% with diarrhea that could not be explained by any other condition and were attributed to VL.25, 26 Moreover, in 16% of the cases, VL was concomitant with another opportunistic infection, a finding previously reported in other studies.7, 8 This association could be explained by the low CD41 lymphocyte count, the slow clinical development of VL, and an even greater immunosuppression induced by Leishmania.27 With respect to the results of the diagnostic tests, there were no significant differences between the first episode and relapses. The IFAT serology has a sensitivity of 7– 55%2, 3, 7, 8, 28 in HIV1 patients with VL, in contrast with its sensitivity of 87–95%29, 30 in immunocompetent individuals. The sensitivity of 36.9% recorded in our study was in agreement with previously published results. Our best results were obtained with the splenic aspiration sample (100% of sensitivity). Although complications occur in fewer than 1% of the cases when this technique is used,31 one of the 10 aspirates was performed on a huge fibrous spleen and splenectomy had to be carried out. The low yield of bone marrow aspirates (63.0%) is not rare since many AIDS patients in advanced stages of the illness present hypoplasic bone marrow and in these cases clinical samples must be cultured.3 With liver biopsy, a sensitivity of 60% was achieved, probably because this was performed postmortem and not according to the standard procedure. Leishmania can be detected in peripheral blood and a sensitivity of 53% is obtained with Giemsa staining17 and 67% with blood cultures.18 In this study, a sensitivity of 42.1% was obtained with blood culture. In spite of the excellent yield obtained with splenic aspirate, we consider the elective diagnostic technique to be the study of bone marrow with parasitic culture since not only VL but also other causes of fever of unknown origin in HIV1 individuals can be diagnosed.32 The results of the efficacy of treatment were not conclusive since drug administration was not random and posttreatment study of the bone marrow was only performed in

441


17.7% of the cases. In our series, pentavalent antimony and amphotericin B performed equally well and were more effective than pentamidine. However, in efficacy studies carried out in India (where there is resistance to antimonials), even low doses of amphotericin B proved more effective than pentamidine33 and antimonials34 but this is not analogous to the situation in southern Europe, where resistance has not been definitively proved. Neither allopurinol nor rHIFNg appeared to improve the condition of the patients, although there were too few patients to reach reliable conclusions. It is probable that rHIFNg is more important in the treatment of relapsed VL in immunocompetent individuals,35 although a good response has also been reported in HIV1 patients.36 After the surprising results initially obtained with liposomal amphotericin B, later studies showed that most patients relapsed in the long-term follow-up.37 In our study, three of the four patients treated with this drug relapsed. There was a high incidence of side effects. Hyperamylasemia is very frequent in immunocompetent patients38 treated with pentavalent antimonials, although severe complications have been reported in immunocompromised individuals.39 Renal failure is a well-known complication of treatment with amphotericin B, but it is noteworthy that the four patients treated with liposomal amphotericin B did not report toxicity. In our HIV1 patients, as previously described,28 VL showed a clear tendency to relapse (27% at six months and 60.6% at 12 months), with the time between the first and second relapse being shorter than the time between the first episode and first relapse (9.3 and 4.7 months, respectively; P 5 0.035). However, the opposite occurs in cases of VL in the immunocompetent population in which no relapse was found in the eight years of the study. Although there are no conclusive prospective studies, use of a secondary life-long prophylaxis, such as that used for other opportunistic parasitic infections in AIDS patients, appears useful. In a recent retrospective study, pentavalent antimony given once a month was effective in the prevention of VL relapses in HIV-infected individuals.40 Given its pharmacokinetic characteristics and simultaneous efficacy against Pneumocystis carinii,41 pentamidine could be good choice. In fact, six patients who were given secondary prophylaxis with parenteral pentamidine after a first relapse of VL have not relapsed after a mean follow-up time of eight months. This finding is significant given that average time to relapse in this group is 4.7 months. With respect to mortality it is noteworthy that 1) 18.5% of the patients with VL died during the first episode; 2) longterm survival was related to their immunocompromised condition since the relative mortality risk was 2.42 times higher in patients that met AIDS criteria at the time of the diagnosis of VL; and 3) survival curves of patients without AIDSdefining criteria when first diagnosed with VL dying during follow-up were similar to survival curves of the control group of patients (those dying of AIDS who never had VL). These points mean that VL behaves in the same way as any other opportunistic infection defining AIDS. Increased mortality was also associated with a decreased number of platelets, with each reduction of 10,000 platelets increasing the mortality risk ratio by 6%. Thrombocytopenia was observed in advanced stages of AIDS, which was prob-

ably due to a decrease in platelet production and splenic capture.42 Moreover, thrombocytopenia and a decreased CD41 count have already been described as predictive of AIDS and death.43 Prolonged splenomegaly secondary to relapsed episodes of VL combined with advanced immunosuppression could be responsible for the association between thrombocytopenia and increased mortality. While a previous or concomitant diagnosis of AIDS at the first episode of VL was a predictive variable of mortality, the CD4 lymphocyte count was not. This could be due to the fact that most patients had similar CD4 lymphocyte counts that were less than 200/mm3 (89.3%). Therefore, as a mortality predictor, clinical immunosuppression (an AIDS diagnosis) seems to be more important than analytical immunosuppression (CD4 lymphocyte count , 200/mm3). Since some researchers have proposed after detailed epidemiologic studies,44 we also suggest that VL behaves as a truly opportunistic infection and should be included in the list of diseases associated with AIDS because it is more prevalent in the HIV1 than in the HIV2 population, it is usually associated with a lymphocyte CD41 count of less than 200/mm3, and because the clinical course of the disease is modified by HIV infection. Acknowledgments: We are grateful to Dr. M. I. Jimenez (Instituto Carlos III de Madrid) for the isoenzymatic characterization of the Leishmania isolates. We also thank the Infectious Diseases Department of the Hospital Ramon y Cajal for invaluable help and daily care of the patients. Authors’ addresses: Rogelio Lopez-Velez, Jose A. Perez-Molina, Antonio Guerrero, and Fernando Baquero, Clinical Microbiology and Infectious Diseases Department, Hospital Ramon y Cajal, Madrid, Spain. Jesus Villarrubia and Luis Escribano, Hematology Department, Hospital Ramon y Cajal, Madrid, Spain. Carmen Bellas, Pathology Department, Hospital Ramon y Cajal, Madrid, Spain. Francisco Perez-Corral, Biostatistics Department, Hospital Ramon y Cajal, Madrid, Spain. Jorge Alvar, Instituto Carlos III, Madrid, Spain. Reprints requests: Rogelio Lopez-Velez, Microbiologia Clinica, Enfermedades Infecciosas, Hospital Ramon y Cajal, Apartado 31057, Madrid 28034, Spain. REFERENCES

1. Fernandez-Guerrero ML, Aguado JM, Buzin L, Barros C, Montalban C, Martin T, Bouza E, 1987. Visceral leishmaniasis in immunocompromised hosts. Am J Med 83: 1098–1102. 2. Montalban C, Martinez-Fernandez R, Calleja JL, GarciaDiaz JD, Rubio R, Dronda F, Moreno S, Yebra M, Barros C, Cobo J, Martinez MC, Ruiz F Costa, J, 1989. Visceral leishmaniasis (kala-azar) as an opportunistic infection in patients infected with the human immunodeficiency virus in Spain. Rev Infect Dis 11: 655–660. 3. Berenguer J, Moreno S, Cercenado E, Bernaldo de Quiros JCL, Garcia de la Fuente A, Bouza E, 1989. Visceral leishmaniasis in patients infected with human immunodeficiency virus (HIV). Ann Intern Med 111: 129–132. 4. WHO, 1994. Acquired immunodeficiency syndrome (AIDS) - Data as of December 1994. Wkly Epidemiol Rec 70: 5–8. 5. WH0, 1995. Report on the Consultative Meeting on Leishmania/HIV Co-Infection 1995. Geneva: World Health Organization, WHO/LEISH/95.35. 6. Desjeux P, 1992. Human leishmaniases: epidemiology and public health aspects. World Health Stat Q 45: 267–275. 7. Rosenthal E, Marty P, Poizot-Martin Y, Reynes J, Pratlong F, Lafeuillade A, Jaubert D, Boulat O, Dereure J, Gambarelli

442

8.

9.

10. 11.

12.

13.

14.

15. 16.

17.

18.

19.

20.

21. 22.

23.

24.

25.


F, Gastaut JA, Dujardin P, Dellamonica P, Cassuto JP, 1995. Visceral leishmaniasis and HIV co-infection in southern France. Trans R Soc Trop Med Hyg 89: 159–162. Medrano FJ, Hernandez-Quero J, Jimenez E, Pineda J, Rivero A, Sanchez-Quijano A, Velez Y, Viciana P, Castillo R, Reyes M, Carvajal F, Leal M, Lissen E, 1992. Visceral leishmaniasis in HIV21-infected individuals: a common opportunistic infection in Spain? AIDS 6: 1499–1503. CDC, 1993. Revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR Morb Mortal Wkly rep 41 (RR-17): 1–19. Park RAA, 1992. European AIDS definition. Lancet 339: 671. Maazoun R, Lanotte G, Pasteur N, Rioux JA, Kennon MF, Pratlong F, 1981. Ecologie des leishmanioses dans le Sud de la France. Contribution a l’analyse chimiotaxonomique des parasites de la leishmaniose viscerale mediterraneenne. A propos de 55 souches isolees en Cevennes, Cote d’Azur, Corse et Tunisie. Ann Parasitol Hum Comp 56: 131–146. Rioux JA, Lanotte G Serres E, Pratlong F, Bastien P, Perieres J, 1990. Taxonomy of Leishmania. Use of isoenzymes. Suggestions for a new classification. Ann Parasitol Hum Comp 65: 111–125. Abraira V, Zaplana J, 1984. PRESTA: Un Paquete de Procesamientos Estadisticos. Gijon, Spain: Proceedings de la Conferencia Iberoamericana de Bioingenieria, 101. Botet J, Portus M, 1993. La Leishmaniosis en la Espan˜a peninsular. Revision historico-bibliografica (1912-1985). Rev San Hig Pub 67: 255–266. Alvar J, 1994. Leishmaniasis and AIDS co-infection: the Spanish example. Parasitol Today 10: 160–163. Marty P, Lelievre A, Quaranta JF, Rahal A, Gari-Toussaint M, La Fichoux Y, 1994. Use of the leishmanin skin test and Western blot analysis for epidemiological studies in visceral leishmaniasis areas: experience in a highly endemic focus in Alpes-Maritimes (France). Trans R Soc Trop Med Hyg 88: 658–659. Martinez P, De la Vega E, Laguna F, Soriano V, Puente S, Moreno V, Sentchordi MJ, Garcia-Auado C, Gonzalez la Hoz J, 1993. Diagnosis of visceral leishmaniasis in HIVinfected individuals using peripheral blood smears. AIDS 7: 227–230. Lopez-Velez R, Laguna F, Alvar J, Perez-Molina JA, Molina R, Martinez P, Villarrubia J, 1995. Parasitic culture of buffy coat for diagnosis of visceral leishmaniasis in human immunodeficiency virus-infected patients. J Clin Microbiol 33: 937–939. Molina R, Canavate C, Cercenado E, Laguna F, Lopez-Velez R, Alvar J, 1994. Indirect xenodiagnosis of visceral leishmaniasis in 10 HIV-infected patients using colonized Phlebotomus perniciosus. AIDS 8: 277–279. Vigilancia Epidemiologica del SIDA/VIH, 1994. Informe no. 14. Madrid: Comunidad de Madrid. Consejeria de Salud. Direccion General de Prevencion y Promocion de la Salud. Alvar J, Jimenez M, 1994. Could infected drug-users be potential Leishmania infantum reservoirs? AIDS 8: 854. Jimenez MI, Ferer-Dufol M, Canavate C, Gutierrez-Solar B, Molina R, Laguna F, Lopez-Velez R, Cercenado E. Dauden E, Blazquez J, Ladron de Guevara C, Gomez J, de la Torre J, Barros C, Altes J, Serra T, Alvar J, 1995. Variability of Leishmania (Leishmania) infantum among stocks from immunocompromised, immunocompetent patients and dogs in Spain. FEMS Microbiol Lett 131: 197–204. Pratlong JP, Dedet JP, Marty P, Portus M, Deniau M, Dedeure J, Abranches P, Reynes J, Martini A, Lefebvre M, Rioux JA, 1995. Leishmania-human immunodeficiency virus coinfection in the Mediterranean basin: isoenzymatic characterization of 100 isolates of the Leishmania infantum complex. J Infect Dis 172: 323–326. Russell DGV, Xu S, Chakraborty P, 1992. Intracellular trafficking and the parasitophorus vacuole of Leishmania mexicana-infected macrophages. J Cell Sci 103: 1193–1210. Matheron S, Cabie A, Parquin F, Mayaud C, Roux P, Antoine

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

M, Chougnet C, Coulaud JP, 1992. Visceral leishmaniasis and HIV infection: unusual presentation with pleuropulmonary involvement, and effect of secondary prophylaxis. AIDS 6: 238–239. Laguna F, Garcia-Samaniego J, Soriano V Valencia E, Redondo C Alonso MJ, Gonzalez la Hoz JM, 1994. Gastrointestinal leishmaniasis in human immunodeficiency virusinfected patients: report of five cases and review. Clin Infect Dis 19: 48–53. Ho JL, Badaro R, Schwartz A, Dinarello CA, Gelfand JA, Sobel J, Barral A, Barral Netto M, Carvalho EM, Reed SG, Johnson WD Jr, 1992. Diminished in vitro production of interleukin-1 and tumor necrosis factor during acute visceral leishmaniasis and recovery after therapy. J Infect Dis 165: 1094–1102. Montalban C, Calleja JL, Erice A, Laguna F, Clotett B, Podzamczer D, Cobo J, Mallolas J, Yebra M, Gallego A, 1990. Visceral leishmaniasis in patients infected with human immunodeficiency virus. J Infect 21: 261–270. Duxburry ER, Sadun HE, 1964. Fluorescent antibody test for the serodiagnosis of visceral leishmaniasis. Am J Trop Med Hyg 13: 525–529. Shaw JJ, Voller A, 1964. The detection of circulating antibody to kala-azar by means of immunofluorescence techniques. Trans R Soc Trop Med Hyg 58: 349–352. Gari-Toussaint M, Lelievre A, Marty P, Le Fichoux Y, 1994. Contribution of serological test to the diagnosis of visceral leishmaniasis in patients infected with the human immunodeficiency virus. Trans R Soc Trop Med Hyg 88: 301– 302. Miralles P, Moreno S, Perez-Tascon M, Cosin J, Diaz MD, Bouza E, 1995. Fever of uncertain origin in patients infected with the human immunodeficiency virus. Clin Infect Dis 20: 872–875. Mishra M, Biswas UK, Jha DN, Khan AB, 1992. Amphotericin versus pentamidine in antimony-unresponsive kalaazar. Lancet 340: 1256–1257. Mishra M, Biswas UK, Khan AB, 1994. Amphotericin versus sodium stibogluconate in first-line treatment of Indian kala-azar. Lancet 344: 1599–1600. Badaro R, Falcoff E, Badaro FS, Carvalho EM, Pedral-Sampaio D, Barral A, Carvalho JS, Barral-Netto M, Brandely M, Silcva L, Bina JC, Teixeira R, Falcoff R, Rocha H, Ho JL, Johnson WD Jr, 1990. Treatment of visceral leishmaniasis with pentavalent antimony and interferon gamma. New Engl J Med 322: 16–21. De Gorgolas M, Castrillo JM, Fernandez Guerrero ML, 1993. Visceral leishmaniasis in patients with AIDS: report of three cases treated with pentavalent antimony and interferon-gamma. Clin Infect Dis 17: 56–58. Davidson RN, Russo R, 1994. Relapse of visceral leishmaniasis in patients who were coinfected with human immunodeficiency virus and who received treatment with liposomal amphotericin B. Clin Infect Dis 19: 560. Gasser RA, Magill AJ Jr, Oster N, Franke DE, Grogl M, Berman JD, 1994. Pancreatitis induced by pentavalent antimonial agents during treatment of leishmaniasis. Clin Infect Dis 18: 83–90. Halim MA, Alfurayh O, Kalim ME, Dammas S, Al-Eisa A, Damanhouri G, 1993. Successful treatment of visceral leishmaniasis with allopurinol plus ketoconazole in a renal transplant recipient after the occurrence of pancreatitis due to stibogluconate. Clin Infect Dis 16: 397–399. Ribera E, Ocana I, Otero J, Cortes E, Gasser I, Pahissa A, 1996. Prophylaxis of visceral leishmaniasis in human immunodeficiency virus-infected patients. Am J Med 100: 496–501. Ena J, Amador C, Pasquau F, Carbonell C, Benito C, Gutierrez F, Vilar A, 1994. Once-a-month administration of intravenous pentamidine to patients infected with human immunodeficiency virus as prophylaxis for Pneumocystis carinii pneumonia. Clin Infect Dis 18: 901–904.


42. Najean Y, Rain JD, 1994. The mechanism of thrombocytopenia in patients with HIV infection. J Lab Clin Med 123: 415–420. 43. Graham NM, Piantadosi S, Park LP, Phair JP, Rinaldo CR, Fahey JL,1993. CD41 lymphocyte response to zidovudine

443

as predictor of AIDS-free time and survival. J Acquir Immune Defic Syndr 6: 1258–1266. 44. Gradoni L, Scalone A, Gramiccia M, Troiani M, 1996. Epidemiological surveillance of leishmaniasis in HIV21-infected individuals in Italy. AIDS 10: 785–791.