Epidemiologic investigation

A total of 39 HIV-seropositive patients were identified by mid-July 1993 in Gharbia. The 39 HIV-infected patients included 34 (61.8%) of 55 patients undergoing dialysis at UCA and 5 (41.6%) of 12 patients undergoing dialysis at PCA. Of the 34 HIV-infected persons undergoing dialysis at UCA, 21 (61.8%) were seroconverters and 13 were seropositive; of the 5 undergoing dialysis at PCA, 3 (60.0%) were seroconverters and 2 were seropositive.

One UCA dialysis patient died before any outbreak data had been collected, leaving for interview the 5 patients from PCA and 33 of 34 HIV-infected patients from UCA. The median age of these patients was 39 years (range, 18–71 years), and 23 (60.5%) were men. In confidential interviews, all 38 denied intravenous drug use, homosexual sex, extramarital sex, or travel to countries with high prevalence of HIV. All first positive HIV tests for these 38 persons were obtained between 1 May and 8 July 1993.

For the 21 seroconverters at UCA, prior negative HIV tests had been performed 21–416 days (median, 240 days) before the first positive test. The number of dialysis sessions ranged from 10 to 305 (mean, 93.9). Eleven of 33 patients with HIV infection had received dialysis previously at PCA. Data available for 10 of these 11 patients revealed that they had had dialysis previously at PCA 1–48 times (mean, 10.9). When seroconversion rates by dialysis day and shift were compared (by use of all patients who were ever seronegative for the denominator), patients who received dialysis on the Saturday/Tuesday schedule at UCA were more likely (12/15 [80%]) to seroconvert than were patients on the Sunday/Wednesday schedule (5/12 [42%]; RR, 1.92; P = .049) or the Monday/Thursday schedule (4/14 [28.6%]; RR, 2.8; P = .005). The rate of HIV seroconversion was higher among patients receiving dialysis on the first (10/14 [71%]) or second shift (8/15 [53%]) than among those receiving dialysis on the third (evening) shift (3/12 [25%]; RR, 2.5; P = .03). HIV seroconversion rates were similar among men and women (14/28 [50%] vs. 7/13 [54%]); P = .82).

Of 20 seroconverters who were known to have been transfused, 12 had a negative test for HIV antibody >30 days after the last transfusion. All but 1 transfusion occurred in Egypt. Some blood units were obtained from paid professional donors of private blood banks. Compared with HIV-seropositive patients, seroconverters at UCA were more likely to have had transfusions (20/21 [95%] vs. 8/12 [67%]; P = .05) or to have had >100 sessions at UCA (13/21 [62%] vs. 1/12 [8%]; P = .003). When controlled for number of sessions at UCA, prior transfusion was still associated with seroconversion, but this relationship was no longer statistically significant (P = .24).

Dialysis center visits

PCA, UCA, center B, and the transplant hospital dialysis center in Gharbia were visited twice during the investigation. They had all been operating for >2 years by May 1993. In 1993, guidelines produced by the Egyptian Ministry of Health in response to the 1990 dialysis-related HIV transmissions in Egypt were in effect. These guidelines recommended HIV screening of all new patients before initiation of dialysis in each center, and standard (universal) precautions to prevent transmission of bloodborne pathogens as outlined in Centers for Disease Control and Prevention guidelines [10]. There was a transfer of patients, but not personnel, between the centers. PCA, UCA, and center B did not have needle or sharps containers, but the transplant hospital dialysis center did have these, as well as red waterproof bags for infectious waste (table 1).

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Figure 1

Amino acid alignment of V3 region from outbreak (boxed, sequences 1–13) and nonoutbreak (sequences 14–21) sequences. V3 loop is in italics. Amino acids that are unique to outbreak sequences are shaded. Note uniformity among outbreak strain amino acid sequences and heterogeneity in this region among nonoutbreak sequences (GenBank accession nos. AF155706-AF155727).

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Figure 2

Phylogenetic tree showing relatedness of outbreak (sequences 1–13) and nonoutbreak (sequences 14–21) sequences from V3 region (238 positions). Aligned sequences were used to generate neighbor-joining phylogenetic tree by means of software package TREE-CON 1.2 and Kimura algorithm. Tree is rooted on sequence of patient 13 (GenBank accession nos. AF155706-AF155727).

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 1

Characteristics of dialysis centers visited in Egypt, June and November 1993.

PCA was located in a poorly maintained private apartment and had 5 dialysis machines (Cobe Century 2; Travenol Laboratories, Deerfield, IL). The total census of patients undergoing dialysis was 5, although there was capacity to dialyze 12–15 patients in 3 shifts. The dialysis staff had no formally trained nurses; training had been given by the medical director of PCA. On a single wheeled tray between 2 patients were unlabeled small (∼25 mL) plastic bottles that were periodically refilled from a larger container (∼1000 mL) said to contain sterile water. Heparin from break-top glass ampules was diluted by drawing up water from these small plastic bottles into unlabeled plastic disposable syringes on the single wheeled tray. As often as every hour, diluted heparinized saline from 1 of the 2 unmarked syringes was administered through the blood line port, and the unmarked syringe was returned to the wheeled tray. Staff members agreed that a single syringe may have been used for >1 patient. In addition, nurses reported that, often, a 10-mL syringe with heparinized saline was carried in their pockets for flushing lines as needed. The staff denied reuse of dia-lyzers (Althin, CD, Miami) or blood lines, and there was no place to reprocess or store reused items. The medical director stated that all patients were tested for HIV on admission to the PCA, and subsequently every 3 months, even though there were no available documents at PCA.

UCA was located within the University Hospital, and its medical director was the same as the medical director of PCA. UCA operated 3 shifts of 6–7 patients who received dialysis twice a week. The UCA staff consisted of professionally trained licensed nurses, 3 of whom worked in each shift. There were 6 dialysis machines. Saline for priming the dialyzers and for dilution of heparin was taken from a sterile normal saline solution bottle dedicated for use for a particular patient. Heparinized saline was kept in 10-mL syringes for boluses for an individual patient on top of the dialysis machine to which the patient was connected. After each use, the CF Baxter dialyzers (Travenol) and tubing were discarded. Reuse was denied, and there were no facilities to reprocess or store used materials.

The UCA staff indicated that before the HIV outbreak was recognized, the same syringe was used to flush the lines of >1 patient. Heparin was diluted with sterile saline in 10-mL syringes prior to or during dialysis sessions. Prefilled heparinized saline syringes were refrigerated overnight. During the dialysis session, they were carried by nurses in their pockets or placed unlabeled on a wheeled tray shared by all shift patients. It is possible that they were used for >1 shift or refrigerated after use.

Center B had 3 machines (2 Cobe [GAMBRO Healthcare, Lakewood, CO] and 1 Fresenius [Fresnius, Bad Homburg, Germany]), and only 1 patient was receiving dialysis during the November visit. Before July 1993, 2 shifts of 2 patients each, 6 days a week, received dialysis in this facility; there was only 1 tray, and flushing with shared syringes may have occurred. The transplant hospital dialysis center had 12 Gambro dialysis machines (GAMBRO Healthcare) and used Baxter 15 hollow-fiber dialyzers (Baxter, Deerfield, IL). There was a separate room for 2 other dialysis machines and 1 nurse dedicated only to hepatitis B surface antigen-positive patients. There were no facilities for reprocessing of dialyzers; reuse was denied. No practices that could result in inadvertent or intentional use of syringes or needles for >1 patient were observed in this center. Used dialyzers and tubing were discarded in infectious waste bags. A syringe of heparinized saline was prepared for each patient at the beginning of the session, labeled, kept on the dialysis machine, and discarded at the end of the session in a rigid puncture-resistant container. There were no patients with HIV infection among the 75 who received dialysis in 3 shifts twice a week at the transplant hospital dialysis center.

Laboratory investigations

Strains related to the outbreak were sequenced; these strains were obtained from 1 dialysis patient whose peripheral blood mononuclear cells had been stored at −130°C since 1994 and from 12 dialysis patients from the 1993 outbreak who were alive in 1995 and 1996 and agreed to donate blood. Two samples were from the same patient (patient 5) obtained at different times, in March and December 1996. In all, 14 samples from 13 patients were available related to the outbreak: 7 patients were HIV seroconverters and 6 were HIV-seropositive patients without prior negative serologic tests (table 2). For comparison, samples from 8 Egyptians whose HIV infection was not related to the 1993 dialysis outbreak (table 3) were sequenced.

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 2

Dialysis history and dates of human immunodeficiency virus (HIV) antibody testing for patients undergoing renal dialysis at 3 dialysis centers in Gharbia, Egypt, and whose strains were analyzed genetically.

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 3

Characteristics of human immunodeficiency virus-positive Egyptian patients not associated with 1993 renal dialysis outbreak whose strains were analyzed genetically.

Of the 14 samples from 13 patients whose samples were related to the outbreak, 7 were from patients who received dialysis only at UCA, and 5 (including the 2 samples from patient 5) were from patients who received dialysis only at PCA (table 2). Another sample was from patient 10, who started dialysis at UCA in February 1992, was transferred to center B early in January 1993, and was confirmed to be HIV seropositive in October 1993. One sample was from patient 8, who never received dialysis at PCA or UCA but did receive dialysis at center B in Gharbia, where patient 10 was undergoing dialysis (table 2). Patient 8, who was transferred to center B in May 1993, was confirmed to be an HIV seroconverter in October 1993.

All outbreak sequences belonged to HIV-1, subtype B. Within the nonoutbreak sequences, 1 strain was subtype E (patient 14), and 2 strains (patients 16 and 19) had quite divergent B sequences with an APGR amino acid motif at the tip of the V3 loop. The nucleotide homology between the outbreak strain sequences was significantly higher (⩾96%) than between sequences of strains unrelated to the outbreak (76%–89%). Alignment of the V3 regions of outbreak and nonoutbreak strain sequences showed that outbreak strain sequences varied much less than did nonoutbreak sequences (figure 1). Two amino acids at positions 6 and 10 after the C-terminal cysteine of the V3 loop (figure 1, shaded Q and A) were unique for the outbreak sequences. A third amino acid at position 14 after the C-terminal cysteine also was unique for the outbreak sequences, except for one nonoutbreak sequence from patient 20.

Two sequences derived from patient 5 from samples obtained 9 months apart in 1996 were included in the phylogenetic analysis. The 2 intraindividual sequences were interspersed in the phylogenetic tree with sequences derived from different dialysis patients, reflecting the clonal relationship of interpatient HIV sequences (figure 2). The homology was remarkable, because the outbreak strains available for analysis for 7 of 12 patients were from blood samples collected in 1996, as much as 3 years after the outbreak. Enough time would have elapsed to promote divergence and development of quasispecies in infecting HIV of each of these patients.

Although patient 8 had never undergone dialysis at UCA or PCA, the HIV strain isolated from patient 8 was closely related to the outbreak strain. This finding suggested that the source of HIV infection for patient 8, who seroconverted in October 1993 while undergoing dialysis at center B, may have been patient 10, who had been transferred from UCA to center B in January 1993.

Control measures

The unexpected finding in early May 1993 of 3 HIV-seropositive dialysis patients during routine periodic screening prompted HIV screening of all dialysis patients in Gharbia. In July 1993, PCA was closed. All HIV-infected patients receiving dialysis at PCA and UCA were transferred to another center designated by the Ministry of Health. The shifts at UCA were reduced from 3 to 1, with 5 HIV-seroneg-ative patients receiving dialysis twice a week. PCA reopened again in October 1993 to dialyze 5 HIV-seronegative patients twice a week in 1 shift. A survey and educational programs in dialysis centers were conducted to emphasize compliance with national guidelines in centers in Gharbia.