Confirmation of Nosocomial Transmission of Hepatitis C Virus by Phylogenetic Analysis of the NS5-B Region

Abstract

Four hepatitis C virus transmission chains at three dialysis units were disclosed by limited sequencing; three of these were disclosed by analysis of the NS5-B region of the genome. Dialysis on the same shift as that during which infected patients were dialyzed was the common factor for seven patients in two chains. Two nurses exposed to needle sticks and their sources of infection constituted two other chains. The strains of three chains belonged to subtype 1a and formed clusters with an intrachain variability of 0 to 6 nucleotides compared to 8 to 37 nucleotides for unrelated strains within this subtype. The clusters were supported by bootstrap values ranging from 89 to 100%.

Hepatitis C virus (HCV) is a major agent of transfusion-transmitted hepatitis (9). Nosocomial spread of HCV has, however, been reported to occur during dialysis (2, 4, 11, 14, 20) and in hematological wards (1), and this explains the high HCV prevalence in nontransfused patients in these settings (16). There are six genotypes of HCV designated by Arabic numerals and an increasing number of subgenotypes or subtypes designated by letters (15, 17). In the Western world genotypes 1, 2, and 3 make up the majority of HCV strains, with genotype 1 being most frequent (17). Within a genotype, sequence variability has been used to characterize strains to identify nosocomial transmission (1, 2, 4, 9, 11, 14, 20) and transmission from batches of immunoglobulins (6, 7, 13) and through needle sticks (19).

We have investigated the association between cases of infection of patients and staff with HCV at three different dialysis units in different towns in central Sweden by sequencing within the NS5-B region (Table 1). Twenty-one other HCV-infected individuals provided strains representative for Sweden (Table 2).

TABLE 1.

Data on patients and genotypes of recovered HCV strains in the three investigations

Patient no.	Sex	Yr born	Origin of serum sample	Epidemiological backgrounda	Date of onset of HCV infection	Serum samplee		HCV genotype
						Date taken	Identification
1:1	M	1946	Karlstad	DP	Dec. 1989b	Dec. 1991	1182/92d	3a
1:2	F	1971	Karlstad	DP	Jan. 1992b	May 1992	7394/93	1a
1:3	F	1921	Karlstad	DP	Aug. 1992c	Aug. 1992	7398/93	1a
1:4	M	1915	Karlstad	DP	Aug. 1992c	Nov. 1992	7400/93	1a
1:5	M	1919	Karlstad	DP	Aug. 1992c	Nov. 1993	316/94	1a
1:6	F	1966	Karlstad	DP	Nov. 1992b	Jan. 1995	286/95	1a
1:7	M	1923	Karlstad	DP	Apr. 1993b	Apr. 1993	7395/93d	1a
1:8	M	1941	Karlstad	DP	June 1994c	June 1994	974/94	2b
1:9	F	1945	Karlstad	Nurse	Oct. 1994b	Oct. 1994	1047/94	3a
2:1	M	1919	Stockholm/Avesta	DP	1990b	Dec. 1995	699/95	1a
2:2	M	1924	Stockholm/Avesta	DP	1990b	Dec. 1995	700/95	1a
2:3	M	1943	Avesta	DP	Nov. 1990b	Oct. 1994	681/95	1a
2:4	M	1972	Avesta	DP	Sep. 1993	Nov. 1995	695/95	1a
2:5	M	1955	Avesta	DP	June 1995	Nov. 1995	694/95	1a
2:6	F	1926	Avesta	DP	June 1995	Nov. 1995	696/95	1a
3:1	F	1934	Uppsala	Nurse	June 1993c	June 1993	6260/93	1a
3:2	M	1923	Uppsala	DP	Dec. 1991b	Jul. 1992	406/94	1a

^aDP, dialysis patient. 

^bFirst known anti-HCV-positive sample. 

^cFirst anti-HCV-negative and HCV PCR-positive sample. 

^dAnalysis of the C (core) region. 

^eFor the HCV strain used for genetic analysis. 

TABLE 2.

Data on Swedish background HCV strains used in the phylogenetic analyses

Serum sample			HCV genotype
Identification	Origin	Background
484/94	Jönköping	Immunoglobulin	1a
586/96	Jönköping	Immunoglobulin	1a
504/94	Halmstad	Immunoglobulin	1a
531/94	Avesta	Immunoglobulin	1a
3817/91	Stockholm	Chronic hepatitis	1a
506/96	Uppsala	Dialysis	1a
906/96	Västerås	Dialysis	1a
911/96	Ängelholm	Sexual transmission	1a
6723/91	Uppsala	Dialysis	1b
7457/91	Karlstad	Blood donor	1b
7458/91	Karlstad	Blood donor	1b
6099/91	Stockholm	Karolinska Hospital	1b
7739/91	Stockholm	Karolinska Hospital	1b
4917/91	Stockholm	Karolinska Hospital	1b
7458/93	Stockholm	Karolinska Hospital	1b
6097/91	Stockholm	Karolinska Hospital	1b
912/96	Ängelholm	Sexual transmission	1b
3453/91	Stockholm	Blood donor	2b
6716/91	Uppsala	Dialysis	2b
1109/94	Uppsala	Dialysis	2b
1041/91	Uppsala	Chronic hepatitis	2b
460/94a	Uddevalla	Sexual transmission	3a
461/94a	Uddevalla	Sexual transmission	3a

^aAnalysis of the C (core) region. 

One investigation comprised 8 of 10 anti-HCV-positive patients at a unit where 60 patients were dialyzed. Patients 1:1 and 1:2 were known seropositives. Since 1991 all patients have been screened for anti-HCV every third month. In November 1992 patients 1:3, 1:4, 1:5, and 1:6 seroconverted. Sera taken in August from patients 1:3, 1:4, and 1:5 were retrospectively found to be HCV RNA positive by PCR. Patients 1:3, 1:4, and 1:5 had been dialyzed in the same room and during the same shift as patient 1:2 in May. There was no earlier serum available from patient 1:6, who had not been dialyzed with patients 1:2, 1:3, 1:4, and 1:5. In April 1993 patient 1:7 seroconverted, and in June 1994 patient 1:8 seroconverted. During September 1994 all staff members were screened for anti-HCV, and a nurse, 1:9, was found to be positive. It then came to light that she had been exposed to patient 1:1 by a needle stick in 1989.

Another investigation comprised all 6 anti-HCV-positive patients at another unit where 43 patients were dialyzed. Patients 2:1 and 2:2, known to be anti-HCV positive, had been dialyzed at this unit during the summer since 1990. In August 1990 patient 2:3 was dialyzed in the same room and during the same shift as patient 2:2. In November patient 2:3 was found to be anti-HCV positive. In July 1993 patients 2:4 and 2:5 were also dialyzed with patient 2:2, and patient 2:4 seroconverted in August. In April 1995 patients 2:5 and 2:6 were dialyzed with patient 2:4, and in June patients 2:5 and 2:6 both seroconverted.

The third investigation dealt with another nurse, 3:1, who in April 1993 was exposed to anti-HCV-positive patient 3:2 by a needle stick. In June she was HCV RNA positive, and in September she seroconverted.

The NS5-B region was chosen for analysis, since it is not considered to be subject to a high level of immune pressure. Viral RNA was extracted from serum mainly as described by Garson et al. (6). cDNA synthesis was performed with Moloney murine leukemia virus reverse transcriptase and primer hep102 (5′-AGCATGATGTTATCAGCTCC-3′) at positions 8681 to 8700 (3). The cDNA was amplified by PCR with hep101 (5′-ATACCCGCTGCTTTGACTC-3′) at positions 8258 to 8276 and hep102. Nesting and sequencing were performed with hep101 and hep105 (5′-ATACCTAGTCATAGCCTCCGTGA-3′) at positions 8616 to 8638. The obtained sequences were aligned with the corresponding parts of 116 HCV sequences from GenBank, and that of the GB virus B (GBV-B) genome (10) by using Tree-Align software (8). A dendrogram was created with DNADIST and NEIGHBOR, PHYLIP package, version 3.53 (5) by using GBV-B as the outgroup. SEQBOOT was used to bootstrap 1,000 data sets.

For some strains the NS5-B region could not be amplified. In these cases part of the core region was amplified with the primers NCR3 (6) and 186 (12). Primers 256 (12) and hep140 (5′-TGAGCACGAATCCTAAACCTCA-3′) at positions 343 to 364 were used for nesting and sequencing. The obtained sequences were aligned with 49 HCV sequences from GenBank. Neighbor joining and bootstrapping were performed as described above.

The genotypes of 17 different HCV strains recovered from the three units are shown in Table 1. Dendrograms based on NS5-B and core regions are shown in Fig. 1 and 2, respectively.

FIG. 1.

Part of a dendrogram, representing the branches formed by genotype 1 strains, obtained by neighbor joining of 300 nucleotides within the NS5-B regions of 156 sequences. The strains recovered in the three investigations are identified by the numbers of the patient from which they were taken (see Table 1). The accession numbers of sequences from GenBank are given at the nodes. Bootstrap values obtained from 1,000 replicas are given as percentages at the branching forks. Boxed strains are those investigated in this study.

FIG. 2.

Dendrogram obtained by neighbor joining based on 256 nucleotides within the C (core) regions of four HCV strains recovered in the investigations, two Swedish background strains, and 49 strains from GenBank. Genotypes are not given, since the analysis did not discriminate between strains belonging to different genotypes. The bootstrap value for the association of strains from patients 1:1 and 1:9 obtained from 1,000 replicas is given as a percentage at the fork formed by these strains. The strains investigated in this study (boxes) are identified by the numbers of the patients from which they were taken (see Table 1).

The HCV strains from patients 1:2, 1:3, 1:4, and 1:5 differed by three to six nucleotides and clustered within subtype 1a with 89% bootstrap support, confirming that patient 1:2 was the source of infection for patients 1:3, 1:4, and 1:5. The strain from patient 1:6 differed from the strains from patients 1:2, 1:3, 1:4, and 1:5 by 14 to 18 nucleotides, while that from patient 1:8 belonged to subtype 2b.

The NS5-B regions of the HCV strains from patients 1:1 and 1:7 could not be amplified. Therefore, the core regions of these strains were sequenced along with those of strains from patient 1:2 and from patient 1:9, the nurse presumed to be infected by patient 1:1. The strain from patient 1:7 was more related to HCV-1 of U.S. origin than to the strain from patient 1:2. Strains from patients 1:1 and 1:9 belonged to subtype 3a and differed by three nucleotides.

All six strains recovered from the second dialysis unit belonged to subtype 1a. Five of these strains from patients 2:2, 2:3, 2:4, 2:5, and 2:6, showed a maximum divergence of three nucleotides and formed a cluster supported by 93% bootstrap. The strain from patient 2:1 differed by 8 to 10 nucleotides from those from patients 2:2, 2:3, 2:4, 2:5, and 2:6.

The HCV strains from the nurse (3:1) and her presumed source of infection (patient 3:2) differed by two nucleotides and clustered with 100% bootstrap support. One of four background strains from this unit belonged to subtype 1a, one belonged to subtype 1b, and two related strains belonged to subtype 2b.

In our investigations four independent chains of HCV transmission were confirmed, two of which involved the transmission of HCV between patients. At one unit four of the strains were closely related, and at a second unit five were closely related, although the strains from different units were unrelated. Comparing the sequences of strains within the same transmission chains showed that the divergence was 0 to 6 nucleotides, while it was 8 to 37 nucleotides for unrelated strains within the same subtype.

That the dialysis patients were the sources of infection for the two nurses confirms that HCV-infected dialysis patients have a virus load compatible with needle stick transmission, which may also explain the frequent horizontal non-transfusion-related transmission of HCV at dialysis centers reported herein and in other articles (2, 14, 18). The exact mode of transmission between the dialysis patients was not clarified. There were no known violations of the infection control precautions, but the finding that transmission resulted from dialysis of noninfected and infected patients on the same shift, although the patients did not use the same dialysis machine, confirmed a previous report of transmission between dialysis patients that also could be related to treatment during the same shift (2).

We have compared only one sequence from each case. HCV strains in dialysis patients have, however, been shown to be heterogeneous by several techniques (14, 18, 20). It may be argued that several HCV strains from each individual should have been sequenced. However, sequencing PCR-generated amplimers provides the consensus sequence for the infected individual, an estimate of which by cloning would have been laborious. In conclusion it was shown that sequencing a PCR fragment within the NS5-B region and phylogenetic analysis of the sequences provided a powerful instrument for investigating nosocomial HCV transmission, despite the fact that most of the strains belonged to the same subtype.

Nucleotide sequence accession numbers.

The nucleotide sequence data reported in this paper will appear in the EMBL, GenBank, and DDBJ nucleotide sequence databases under accession no. AF071953 to AF071987.