Citation analysis for measuring the value of scientific publications: quality assessment tool or comedy of errors?

The basic concepts of citation analysis and journal impact factors are discussed in the light of quality assessment of scientific publications, individual scientists and research units. The major controversies concerning this topic are addressed: technical limitations, database selectivity, time and discipline-related biases, language and publication type biases, multiple authorship merits and citing motivations. Both positive and negative aspects are put into perspective. The authors conclude that citation analysis, even when based on journal impact factors, can be a worthwhile criterion for evaluating publication records of individual scientists or research units, as long as some of the problems discussed are sufficiently taken into account. However, this conclusion in no way implies that citation analysis may be considered as the one and only evaluation criterion.

Which health journals are most useful?

Records are kept in a database of the sources cited in publications currently being received at the library of the Institute of Tropical Medicine in Antwerp in the subject areas of health services and endemic diseases in developing countries. The frequency of citations shows which journals are most useful on these two subjects. Although such an analysis necessarily reflects the needs of the institution carrying it out, it helps to form a more complete idea of what is available.

Loa loa: immunological responses during experimental infections in mandrills (Mandrillus sphinx).

Six intact, adult mandrills (Mandrillus sphinx) were infected with human-derived, diurnal Loa loa infective larvae. Microfilaremia, hematological, and immunological parameters were followed for 2-4 years. A major aim was to investigate the relationship between specific humoral immunity to microfilariae and microfilaremia and also to assess whether infection led to generalized immune dysfunction. Microfilaremia was similar to previous studies for 4 mandrills, with a prepatent period of 153.5 +/- 10.1 days, a peak of 34-1,798 mf/ml around Day 200, followed by a decline to low, persistent microfilaremia. One mandrill (No. 20) had a longer prepatent period and very low, but persistent, levels of microfilaremia, and one (No. 19) had gradually increasing levels which remained > 10,000 mf/ml for 3 years. To assess generalized immune perturbations several parameters were studied. There was neither generalized leukocytosis nor relative or absolute eosinophila. Serum Ig concentrations were measured from 0-600 days postinfection by radial immunodiffusion using a rabbit anti-mandrill Ig serum, and these were remarkably stable. Proliferative responses of peripheral blood mononuclear cells from these infected mandrills and noninfected controls showed no significant differences in the magnitude of proliferation after stimulation with a range of doses of PHA, PWM, or Concanavalin A. Thus, no evidence of generalized immune dysfunction was found in the peripheral blood. Serum IgG levels to soluble mf antigens were estimated by an indirect ELISA and all animals had maximal levels around Week 22 postinfection, at the time of maximum microfilaremia, and these decreased over the next 2-3 years in all mandrills except No. 19, in which levels remained fairly constant. Serum IgG levels to adult worm antigens showed a similar pattern but were not, or were only slightly, diminished late in infection. Antibody to mf sheath antigens were detected by indirect immunofluorescence and agglutination of live mf. Antibody to sheath antigens were never detected in mandrill 19 but were present from Week 8 postinfection to 2-4 weeks before patency in all the others. Anti-sheath antibodies were not detected in serum at later time points, i.e., postpatency, in any mandrill, even at time points when microfilaremia was < 1 mf/ml. The anti-sheath antibody was IgM and no anti-sheath IgG was detected. Ig was detected on the surface of circulating mf in 1 mandrill (No. 20). The appearance of these antibodies prior to maturation of the adults indicates that certain L4 or immature adult antigens cross-react with the surface of mf.(ABSTRACT TRUNCATED AT 400 WORDS)

Genetic comparison of HIV-1 isolates from Africa, Europe, and North America.

PIP: Laboratory scientists used anchored polymerase chain reaction (PCR) and DNA sequencing to compare HIV-1 isolates from countries in Africa (Ivory Coast, Gabon, Zaire, Kenya, and others), Europe (Belgium and other countries), and the US. The US isolates had the most homogenous PCR profile followed by the European pattern. There was considerable PCR primer mispairing for the African isolates, especially those from Kenya, indicating that the range of HIV-1 variation could have been rather extensive. This virus diversity could greatly affect therapy or intervention in sites in Africa with such a complex mix of variants. Nevertheless, the genetic information of these diverse isolates could bring about research leading to an anti-HIV-1 vaccine. For example, the expanded DNA sequence data base could record phylogenetic relationships, thereby, helping researchers choose prototypic variants for vaccine development. More information would allow researchers to generate new PCR primers for better discrimination of variants. They could apply PCR typing to huge sample sizes to adequately document HIV-1 variation in Africa. It could also prove invaluable as a means to determine incidence and prevalence of local variants during vaccine field trials. It can also discern the limiting criteria for HIV-1 genetic variation.^ieng

Genetic organization of a chimpanzee lentivirus related to HIV-1.

Simian immunodeficiency viruses have been isolated from four species of monkey, the 'captive' macaque and mangabey and the 'feral' African green monkey and mandrill. While none of these viruses is a replica of HIV-1, the macaque and mangabey viruses represent correct genetic models for HIV-2, possessing exactly the same complement of genes. Recently a lentivirus has been identified in two wild chimpanzees (Pan troglodytes troglodytes) in Gabon, west equatorial Africa, and isolated from one of them. This virus is referred to as SIVCPZ. Sera from these animals cross reacted with all the HIV-1 proteins including the envelope glycoproteins. Here, we describe the molecular cloning and sequencing of an infectious proviral clone of SIVCPZ. The overall genetic organization was the same as that of HIV-1, but phylogenetic analysis revealed that the sequence was more divergent than any HIV-1 sequence reported so far. The vpu gene product, found only in the type 1 viruses, was particularly different (64% divergent to HIV-1BRU) suggesting that the SIVCPZ represents a distinct subtype. These findings indicate that there is a larger pool of simian lentiviruses than previously suspected and revives debate as to the origins of HIV-1.

A highly defective HIV-1 strain isolated from a healthy Gabonese individual presenting an atypical western blot.

In central equatorial Africa the frequency of uninterpretable or atypical Western blots (WB)--ie. antibodies to gag proteins only--can represent up to 50% of enzyme-linked immunosorbent assay (ELISA)-positive samples. To date the significance of such serology remains unknown. Nevertheless, an unusual HIV-1 strain has been isolated from the blood of a healthy Gabonese individual who presented an atypical WB. This virus, identified as isolated HIV-1OYl, grew to low titres of reverse transcriptase activity (less than 50,000 cpm/ml) and was not obviously cytopathic. Radioimmunoprecipitation and peptide ELISA studies indicated that the lack of env-specific reactivity was probably due to the absence of antibodies to the viral glycoproteins, rather than the virus encoding a highly divergent envelope protein. Molecular cloning and sequencing of the provirus proved it to be a string of HIV-1 which was genetically closer to European and North American than to African strains. Furthermore the envelope protein sequence contained all the features of a typical HIV-1 env gene. However, the tat gene derived from the proviral clone was functionally defective. Site-directed mutagenesis of this gene showed that this was due to the substitution of an essential cysteine residue for a serine. Polymerase chain reaction amplification of the tat gene, as well as parts of the gag and env gene sequences of HIV-1OYl, showed that essentially all of the proviruses were defective. These data emphasize the need to view HIV isolates as populations of distinct genomes capable of complementing each other.

Hypogonadism and ecdysteroid production in Loa loa and Mansonella perstans filariasis.

Possible endocrinological repercussions of infection with Loa loa and Mansonella perstans filariae were studied in Gabonese subjects. Microfilaremic males were compared with amicrofilaremic controls. In the infected group 13/105 subjects (12%) presented only abnormally low serum levels of testosterone (less than 4 ng/ml), 25/105 (24%) only abnormally high serum levels of gonadotrophins, FSH (greater than 15 mIU/ml) and LH (greater than 20 mIU/ml), and 22/105 (21%) presented anomalies in both testosterone and gonadotrophin levels. One out of 68 control subjects had 3.6 ng/ml seric testosterone and all had normal levels of gonadotrophins. Ecdysteroids were detected (greater than 0.025 ng/ml) in the serum of 87/97 (90%) microfilaremic subjects (GM 0.123 ng/ml) compared to 12/64 (19%) controls (GM 0.030 ng/ml). Ecdysteroids were detected in the urine of all subjects, infected (GM 8.468 ng/ml) as well as control (GM 1.245 ng/ml). The hormonal perturbations were correlated with the levels of Loa loa microfilaremia but not with those of serum and urinary ecdysteroids. These results demonstrate that microfilaremic subjects often show endocrinal signs of hypogonadism and present appreciable levels of ecdysteroids in serum and urine. A direct role for parasitic ecdysteroids in hypogonadism remains to be demonstrated.

Isolation and characterization of simian immunodeficiency virus from mandrills in Africa and its relationship to other human and simian immunodeficiency viruses.

Two isolates of simian retrovirus related to the human immunodeficiency virus (HIV) were obtained from apparently healthy mandrills, Papio (Mandrillus) sphinx, in western equatorial Africa. This virus, designated SIVMND (simian immunodeficiency virus from mandrills), appeared morphologically similar to HIV by electron microscopy, showed Mg2+-dependent reverse transcriptase activity, and induced cytopathic effect in human CD4-positive cells. Western blotting (immunoblotting) analyses revealed that the gag and pol products of SIVMND showed cross-reactivity with those of known HIVs and SIVs. Molecular clones covering full-length viral DNA were obtained from closed circular extrachromosomal DNA of SIVMND-infected cells. By clone-on-clone hybridization with known retroviruses of the HIV and SIV groups, SIVMND showed similar cross-hybridization with HIV-1, HIV-2, SIVAGM (African green monkey-derived SIV), and SIVMAC (rhesus macaque-derived SIV) in the gag and pol regions only at low stringency but not at high stringency, a result indicating that SIVMND is a new member of the HIV-SIV group. The existence of distinct SIVs in different monkey species suggest that recent interspecies transfer of HIV-SIV is unlikely in nature.

Trypanosoma brucei: polyamine oxidase mediated trypanolytic activity in the serum of naturally resistant cattle.

Trypanosoma brucei brucei are lysed when incubated in vitro in a mixture of bovine serum and polyamine. Normal bovine serum alone or polyamine alone does not show any trypanocidal activity. The bovine serum in the mixture can be replaced by purified polyamine oxidase, and addition of polyamine oxidase inhibitors blocks trypanolysis. Using this in vitro lysis test, it is shown that West African cattle which are resistant naturally to trypanosomiasis have a higher trypanolytic activity in their serum than do trypanosensitive cattle (P less than 10(-5]. Seric trypanolytic activity of individual animals remains stable when tested over a period of 18 months; moreover, it is not modified by trypanosome infection. Higher levels of seric polyamine oxidase in resistant cattle were demonstrated also by enzymatic analysis. The factors responsible for trypanolysis have been analyzed. Oxidation of spermidine by polyamine oxidase leads to the production of unstable aldehydes, acrolein, ammonia, O2-, HO, and H2O2. Acrolein and H2O2 show strong trypanolytic activity while the other products do not appear to be toxic for trypanosomes. The physiological importance of polyamine oxidase mediated trypanolysis is unclear; even at peak parasitemia in cattle (10(7) organisms/ml) it can be calculated that trypanosomes would not release enough spermidine for the generation of sufficient quantities of toxic degradation products. Additional polyamines could be released in serum from tissues damaged as a result of the infection.

Trypanosoma congolense: lack of correlation between the resistance of cattle subjected to experimental cyclic infection or to field challenge.

Twelve male cattle of the Baoulé breed were exposed to natural trypanosome challenge in an area of high Glossina density, to characterize them as trypanoresistant or trypanosensitive. Weekly blood samples were taken for the determination of parasitemia and packed cell volume, as a measure of anemia. Seven Zebu cattle were also exposed to challenge at the same time. The Zebu proved to be trypanosensitive with high parasitemia, pronounced anemia and died or were drug treated in extremis. Five Baoulé were as sensitive as the Zebu while 7 others were trypanoresistant since they showed little or no patent parasitemia, only mild transient anemia and survived in good condition. The 12 Baoulé were allowed to recover from challenge in the field and along with 7 Zebu were subjected to experimental fly challenge in fly-proof accommodation. Glossina morsitans submorsitans infected with a clone of Trypanosoma congolense derived from the stock Serengeti/71/STIB/212 were allowed to engorge on the shaven flanks of tranquilized animals. All animals showed persistent parasitemia for at least 7 weeks, including all the Baoulé resistant to natural challenge. Two Baoulé, one resistant and one sensitive to natural challenge, and 4/7 Zebu appeared unable to control parasitemia, had severe anemia, and were drug treated in extremis. The remaining Baoulé, 6 resistant and 4 sensitive, appeared to be undergoing spontaneous cure by Week 9-10, as did 3/7 Zebu. In Zebu, anemia was as pronounced as under natural challenge. Three resistant Baoulé maintained packed cell volume above 30 as under field challenge but the others showed marked anemia. On the contrary, 4 sensitive Baoulé showed only slight anemia after artificial fly challenge.(ABSTRACT TRUNCATED AT 250 WORDS)

The role of antibody in natural resistance to African trypanosomiasis.

The nature and usefulness of trypanoresistance in West African Baoule cattle was studied by exposing animals in areas of high Glossina density. Under such conditions, Zebu and some Baoule died soon with high parasitaemia while resistant Baoule showed little patent parasitemia, almost no anaemia and thrived. Subsequently the role that specific antibody may play in such trypanoresistance was analyzed. In the first instance we examined protective antibody titres during T. congolense infection of inbred mice strains, comparing results obtained in trypanosome-mouse combinations where the host controls parasitemia and survives with those obtained when the host fails to control parasitemia and dies. We then attempted to extend these observations to cattle by following the disease course and appearance of neutralizing antibodies in animals of known sensitivity to natural Glossina challenges, following artificial challenge with T. congolense infected Glossina.

The virulence of Trypanosoma congolense can be determined by the antibody response of inbred strains of mice.

Three inbred strains of mice, BALB/c, C57Bl/6 and CBA/J were infected with three clones of Trypanosoma congolense, DIND/3.1, SAM/28.1 and KAR/57.1, which were obtained from three different stocks. DIND/3.1 was of high virulence for BALB/c and CBA/J but of negligible virulence for C57Bl/6. SAM/28.1 was of high virulence and KAR/57.1 of negligible virulence for the three strains of mice. In each case, high virulence was correlated with a late, transient and low titre protective antibody response measured by complement mediated lysis of live organisms. Negligible virulence was correlated with an early, high titre protective antibody response. Suppression of the antibody response by sub-lethal irradiation or cyclophosphamide treatment of the host turned a trypanosome infection of negligible virulence into one of high virulence. In mice with mixed infections it was shown that highly virulent trypanosomes did not influence the course of infection and antibody response to trypanosomes of negligible virulence and vice-versa. The relationship of total antigen mass to the kinetics of the antibody response suggests that 1000- to 10,000-fold less antigen is required in good responder than in bad responder mice to trigger the immune response. Thus the virulence of T. congolense can be determined by the antibody response of inbred strains of mice. The specificity and dose dependency of this antibody response seem to implicate the involvement of Ir genes.

Identification and selection of cattle naturally resistant to African trypanosomiasis.

Cattle were exposed to natural trypanosome challenge in an area of high Glossina density (Samandéni, Burkina Faso) for various periods of time during 1982, 1983, 1984 and 1985. All of 30 Zebu proved to be sensitive to trypanosomiasis i.e. they died or were treated in extremis in 10 +/- 4 weeks. Twenty-one (31%) Baoulé were as sensitive as the Zebu while 47 (69%) were resistant i.e. they survived in good condition. Twenty Ndama/Baoulé crosses, indigenous to Samandéni were all resistant. Weekly blood samples were taken (2,317 in total) for the determination of parasitaemia and packed cell volume (PCV) as a measure of anaemia, the most important pathological feature of cattle trypanosomiasis. In both Zebu and sensitive Baoulé 59% of the blood samples showed positive parasitaemia, of which 38% and 52% respectively were T. congolense the major cattle pathogen in the area considered. In resistant Baoulé and Ndama/Baoulé 11% and 10% of the samples were positive for trypanosomes of which only 4% and 2% were T. congolense respectively. PCV decreased from 35 to 20 in Zebu, 39 to 20 in sensitive Baoulé and 40 to 34 in resistant Baoulé, there was no change in the indigenous Ndama/Baoulé. Six Ndama/Baoulé indigenous to Samandéni remained resistant to trypanosomiasis when moved to another area of high Glossina challenge. Seven Ndama/Baoulé calves, conceived in Samandéni but born and kept for 2 1/2 years in a Glossina-free area, also proved to be resistant to challenge. Twelve Baoulé calves, born from cattle selected under natural field challenge, and which had not come in contact with trypanosomes for the first 10 months of their life, proved to be resistant when exposed in the field. These observations show that some, but not all, cattle from the Baoulé breed are naturally resistant to African trypanosomiasis, that this resistance does not need repeated exposure to trypanosomes early in life but appears to be inherited and functional against many types of antigenically different trypanosomes. Thus, selective breeding of trypanoresistant animals and their successful introduction, without trypanocidal drug protection, into areas of high Glossina density appears feasible.

Incidence of bovine leukemia virus-specific antibodies in West African cattle.

A sero-epidemiological survey of bovine leukemia virus (BLV) infection in indigenous West African cattle was undertaken using a radioimmunoassay and enzyme-linked immunosorbent assay. A high incidence of anti-BLV antibodies was found in the breeds tested and their crosses, whatever the test used. More extensive studies are needed to establish the prevalence of BLV infection in other parts of Africa; these may provide additional data on the similarities found between human and bovine leukemia viruses.

Susceptibility and immune responses of zebu and taurine cattle of West Africa to infection with Trypanosoma congolense transmitted by Glossina morsitans centralis.

Following tsetse-transmitted infection with Trypanosoma congolense, major differences in development of localised skin reactions, the ability to control parasitaemia, the degree of anaemia and in antibody response to trypanosomes were found between the reputedly trypanotolerant breeds of cattle (N'Dama, N'Dama/Baoule crosses, Baoule) and the trypanosusceptible West African Zebu. The local skin reactions that developed in the Zebu were large and severe while those that occurred in the other breeds were smaller and less severe or mild. The timing of appearance of parasitaemia and the height of the first peaks were similar in all the animals, but the Zebu were less able to control subsequent waves of parasitaemia. Possibly reflecting these events, it was only in the Zebu that significant anaemia developed. Neutralizing antibody against homologous metacyclic trypanosomes developed between 14 to 18 days after infection in all breeds of cattle; however, marked differences were found when antibody to trypanosomes derived from first peak parasitaemias were tested in the Zebu and Baoule. Neutralizing antibody against these parasites appeared in the Baoule on day 24 but were not detected in Zebu until day 51. Furthermore, the antibody titres were 3 log2 higher in the Baoule. It was concluded that the trypanotolerance exhibited by the West African taurine cattle might be related to a) their ability to control trypanosome numbers in the skin and in the bloodstream, an outcome that was possibly brought about by the earlier and superior immune response and b) failure to develop anaemia which might be associated with their capacity to control parasitaemia.