Didanosine (ddI) use has been consistently associated with the development of non-cirrhotic portal hypertension (NCPH) [1–5]. We read with interest the recent article by Merchante et al. [6] showing an association of ddI use and abnormal liver stiffness (>7.2 kPa) determined by transient elastography (TE). This article raised the important question as to whether there would be benefit from screening individuals exposed to ddI with TE. We performed a retrospective analysis of all individuals within our HIV cohort who had ever received ddI in order to identify the risk of developing this condition with cumulative ddI use and suitability of screening. A case notes review was performed for each individual and those with radiological or endoscopic features of portal hypertension were identified. Individuals with a recognized cause of liver disease were excluded. A χ2 test with trend analysis was performed to assess risk of developing portal hypertension with cumulative ddI in this cohort. A timeline was created to see the time after initiation of ddI to two consecutive pathology results outside the normal range (those with abnormal results at time of initiation of ddI were excluded) or features of portal hypertension on radiology or endoscopy. Within our cohort, 2,070 individuals had been exposed to ddI. Of these, 26 subjects had idiopathic portal hypertension. A χ2 test with trend analysis revealed the development of portal hypertension was significantly associated with cumulative ddI exposure (P=0.0001; Table 1). There were no identifiable cases of portal hypertension in those exposed for <2 years to ddI. Of 26 subjects, 24 had complete data for dates of commencing and stopping ddI. This timeline revealed the median duration of ddI exposure in this group was 66 months (IQR 47–90). Alanine aminotransferase (ALT) was the first parameter to become abnormal after median 32 months of ddI exposure (IQR 18–55) in 22/23 subjects, then alkaline phosphatase (ALP) at median 44 months (IQR 30–62) in 22/23. Platelets fell below the Letter

Non-cirrhotic portal hypertension (NCPH) has been associated with didanosine (ddI) exposure. We aimed to determine the number of individuals with NCPH within our cohort and define their characteristics. We identified individuals within our cohort with NCPH and performed a retrospective case note review. Cumulative antiretroviral therapy (ART) use was calculated and a statistical analysis performed to compare exposure to the rest of the clinic cohort for the same time period. Where available, data was collated on FibroScan®, echocardiography and coagulation profile. Seventeen patients were identified. Upper gastrointestinal bleeding was the most common presenting feature. Liver biopsy showed mild portal or periportal fibrosis in 13 (81%) and four with features of nodular regenerative hyperplasia. There was significantly greater exposure to ddl in this group (59.5 months) compared to the rest of the HIV cohort (21.1 months) P = <0.001. Eleven subjects has a liver elastography performed, six (55%) had a result greater than 9.6 kPa (consistent with greater than F2 disease by Metavir scoring). Echocardiography was performed in seven patients: four met criteria for pulmonary hypertension. This is consistent with other cohorts demonstrating an association between the didanosine exposure and NCPH. Our data also suggest an increased risk of pulmonary hypertension.

We read with interest the recent article from Waters et al. comparing response to antiretroviral therapy (ART) in late presenters (individuals diagnosed and starting ART with a CD4 count <200 cells/μL) and late starters (individuals diagnosed with a CD4 count >350 cells/μL but starting ART at <200 cells/μL) [1]. The article revealed that 3688 individuals commenced ART with a CD4 count <200 cells/μL; of these, 2741 (74%) were deemed late presenters. In their analysis, the majority of clinical events (AIDS-defining illness or death) occurred in late presenters. In contrast, we had noticed that an increasing number of new opportunistic infections (OIs) and AIDS events were occurring in patients with established HIV infection in our cohort. To test the validity of this observation, we performed a review of our cohort to assess whether those presenting for the first time with a serious OI had previously undergone an HIV test. For this analysis, a late presenter was defined as an individual having a reactive HIV test within 6 months of diagnosis of an OI. We performed a retrospective analysis to identify individuals presenting with a new diagnosis of cryptococcal meningitis (CM), cerebral toxoplasmosis or Pneumocystis jirovecii pneumonia (PCP) from 1 January 2005 to 31 December 2010 via electronic clinical codes. We then carried out a case-based notes review to determine HIV test results prior to the diagnosis of an OI and the CD4 cell count and HIV-1 RNA at admission. Data were included for individuals with CM on the basis of a positive cerebrospinal fluid (CSF) culture, PCP on the basis of positive immunofluorescence or high clinical suspicion based on radiology and oxygen desaturation on exercise, and toxoplasmosis on the basis of compatible radiology with response to treatment. Where subjects had more than one admission per OI, data were collected only for the primary presentation. During this time period, 117 serious OIs occurred: nine cases of CM, seven cases of toxoplasmosis and 101 cases of PCP. The median CD4 count was 52 cells/μL [interquartile range (IQR) 18–142 cells/μL] and the median HIV-1 RNA was 84 000 HIV-1 RNA copies/mL (IQR 2696–197 000 copies/mL). Seventy-three individuals (62%) had previously undergone a positive HIV test more than 6 months prior to the diagnosis of an OI and were aware of the result. In these individuals, the median duration from diagnosis of HIV infection to presentation of the OI was 8.5 years (IQR 4.5–13 years). Within this group, 44 of the 73 individuals (60%) had previously commenced ART prior to diagnosis of the OI, and had been on ART for a median of 8 years (IQR 4.5–10.7 years). Our findings also raise the issue of chemoprophylaxis in patients who would not necessarily receive it according to consensus guidelines. None of the individuals diagnosed with toxoplasmosis or CM was on ART; however, seven individuals (7%) with PCP had undetectable HIV-1 RNA, and more details of these patients are given in Table 1. Three of these patients had a CD4 count >200 cells/μL (>15%) and would not routinely be on prophylaxis. The Opportunistic Infections Project Team of the Collaboration of Observational HIV Epidemiological Research in Europe (COHERE) showed, in patients discontinuing PCP prophylaxis after starting ART, that the incidence of primary PCP was zero cases per 1000 person-years of follow-up in those with a CD4 count of 101–200 cells/μL [2]. Four of seven patients were within this category and had an undetectable viral load for a median of 55 months (range 15–75 months). These data show that, even in the era of effective ART, the majority of individuals presenting with serious OIs in our cohort had already received a diagnosis of HIV infection and were not late presenters. There are a multitude of reasons why these individuals present with serious OIs, including poor compliance with treatment, defaulting from follow-up, substance abuse, denial of diagnosis and inadequate prophylaxis. Because of the retrospective nature of this study, individual patient attitudes towards their health could not be explored, but future research in this area may be beneficial. In conclusion, our results highlight the importance of not only starting ART in a timely fashion but engaging the diagnosed population with services and providing ongoing adherence support. In the era of increasing financial restraint, we may need to focus more on our existing patients than on large-scale, low-yield testing strategies.