RESEARCH LETTERS 4 Dabelea D, Hanson RL, Bennett PH, et al. Increasing prevalence of type 2 diabetes in Pima Indian children. Diabetes 1998; 47 (suppl): A25. 5 Expert Committee on the Diagnosis and Classificiation of Diabetes Mellitus. Report of the expert committee on the diagnosis and classification of diabetes millitus. Diabetes Care 1997; 20: 1183–85. Department of Pediatrics and Community Health Sciences, University of Manitoba, and the Island Lake Tribal Council, Winnipeg, Manitoba (H J Dean) 1524 THE LANCET • Vol 352 • November 7, 1998 Detection of an erythrovirus sequence distinct from B19 in a child with acute anaemia Quang Tri Nguyen, Christophe Sifer, Véronique Schneider, Françoise Bernaudin, Véronique Auguste, Antoine Garbarg-Chenon Erythrovirus (formerly parvovirus) B19 causes a wide range of diseases including acute anaemia due to aplasic crisis. Diagnosis of B19 infection is by serology and detection of viral DNA by PCR or DNA hybridisation. These techniques are usually thought to detect all erythrovirus field isolates, because the B19 genome undergoes few genetic variations (<1% in the entire genome). 1 In May, 1995, a 6-year-old child was admitted to a paediatric unit with a fever associated with severe microcytic aregenerative anaemia and lymphoneutropenia. Investigations of haemolysis (haptoglobin <0·09 g/L) showed a glucose-6- phosphate dehydrogenase defect. A B19 infection was sought. B19 serology was IgM negative and IgG positive initially and 3 months later. A PCR search for the B19 viral genome was done on bone marrow and serum samples; results were inconclusive. PCR products of the same size as the B19- positive control were obtained, but failed to hybridise with a Raloxifene-associated hepatitis Antonio R Vilches, Victor Pérez, Dora Eva Suchecki Raloxifene is the first therapeutically available selective oestrogen receptor modulator and is marketed in many countries. A postmenopausal white woman aged 49 years was started on raloxifene (60 mg) and calcium carbonate because of worsening bone-mineral-density values. Liver-function tests were normal before the drug was started. 30 days after starting treatment she complained of progressive malaise, preceded by an upper-body rash, and she had stopped the two drugs. She first presented 5 days after stopping medication with jaundice. Physical examination was otherwise unremarkable, except for a mild, morbilliform rash. She was not pyrexial before nor after assessment. Laboratory tests showed: aspartate aminotransferase 224 U/L, alanine aminotransferase 291 U/L (normal range 0–35 U/L); alkaline Multiple alignments of part of the VP1u gene 346 bp sequences (position +1 corresponds to ATG of VP1 u) are identified by their mnemonics in GenBank. Only sequences from V9, pvbaua (B19 princeps isolate) and most divergent B19 sequences are represented for clarity. B19-specific probe. PCR products were directly sequenced. When compared by FASTA analysis to sequences stored in GenBank and EMBL databases, this 346-bp sequence resembles the B19 unique region of VP1 (VP1u) gene sequences. However phylogenetic analysis showed the sequence from this new isolate (which we call V9) to be unexpectedly more divergent from 24 B19s (11·07% divergence) than these B19 sequences among themselves (6·65% divergence), as illustrated by multiple-sequence alignment (figure). And like the B19 sequences, the V9 sequence was situated far from the animal parvovirus sequences. The simian parvovirus sequence 2 was the closest to V9 with 44% divergence (as with B19). The sequence divergence that we saw between V9 and B19 seems to extend beyond the VP1u region as V9 DNA patterns obtained with the restriction enzymes BamHI, HindIII, and PvuII differed markedly from those of 65 B19 strains of different geographic origins reported in the literature. 3–5 Such high divergence may interfere with diagnosis of erythrovirus infection by B19 PCR, DNA hybridisation, or for B19 serological assays. Diagnostic tests specific to V9 DNA and antibody detection are currently being developed to address the problem. Because B19 IgG serology was initially positive, it appears that anti-B19 antibodies fail to assure cross-protective immunity against V9. The taxonomic position of V9, as a new genotype in the B19 species or a new species in the genus Erythrovirus, remains to be established. Cloning and sequencing of the whole V9 viral genome may help to fully elucidate its taxonomic position. Is V9 an emerging human virus? Epidemiological studies with V9-specific assays may provide some clues. 1 Hicks KE, Cubel RC, Cohen BJ, Clewley JP. Sequence analysis of a parvovirus B19 isolate and baculovirus expression of the non-structural protein. Arch Virol 1996; 141: 1319–27. 2 Brown KE, Green SW, O’Sullivan MG, Young NS. Cloning and sequencing of the simian parvovirus genome. Virology 1995; 210: 314–22. 3 Morinet F, Tratschin JD, Perol Y, Siegl G. Comparison of 17 isolates of the human parvovirus B19 by restriction enzyme analysis. Arch Virol 1986; 90: 165–72. 4 Mori J, Beattie P, Melton DW, Cohen BJ, Clewley JP. Structure and mapping of the DNA of human parvovirus B19. J Gen Virol 1987; 68: 2797–806. 5 Umene K, Nunoue T. The genome type of human parvovirus B19 strains isolated in Japan during 1981 differs from types detected in 1986 to 1987: a correlation between genome type and prevalence.J Virol 1990; 71: 983–86. Laboratoire de Virologie, Hôpital Armand Trousseau, Paris, France (Q T Nguyen; e-mail nqt@ pasteur.fr); Laboratoire de Virologie, Hôpital Rothschild, Paris; Service de Pédiatrie, Centre Hospitalier Intercommunal de Créteil, Créteil