Acute kidney failure more common in people newly diagnosed with HIV

A study from King’s College Hospital in London has found that acute renal failure (ARF) is a common development in patients new to HIV care, within the first three months of diagnosis.

During an eight-year period, 5.7% of patients developed the condition at some point, but the risk of ARF was more than ten times higher during the first three months of HIV care (‘early onset’ ARF) than it was after the sixth month.

Late-onset kidney failure, which affected no more than 1.6% of patients a year after the first six months in care and declined further over time, affected different patients and appeared to be due to different factors.

The condition was severe enough to require dialysis in 29% of patients with renal failure who were newly diagnosed with HIV, and 17% of other patients with HIV, and death occurred within 90 days of the ARF episode in about a third of patients.

This high mortality, however, is not attributed directly to ARF by the authors, and was usually reversible. It appeared to be caused by a number of factors, including AIDS-related and non-AIDS related disease, drugs used to treat them that were nephrotoxic (harmful to the kidneys), and in some cases immune reconstitution inflammatory syndrome (IRIS).

HIV drugs were only occasionally implicated in ARF and only in late-onset kidney failure, more than three months after diagnosis. After indinavir was discontinued as a commonly-used HIV drug, ARV-related kidney toxicity became a rare event.

The study analysed all 2274 patients seen for HIV care at King’s HIV clinic between January 1998 and December 2005. Seventy-seven per cent of them took HAART during the study period.

Acute Renal Failure was defined as a rapid and usually reversible decline in the Glomerular Filtration Rate (GFR) in the context of previously normal kidney function, or in patients with pre-existing chronic kidney disease, in which case it was called ‘acute on chronic’ renal failure.

GFR is the rate at which kidneys filter blood, and ARF was defined as a decline to below 60 millilitres a minute and either a greater than 40% reduction in the pre-existing GFR or serum creatinine reduced more than 50% from baseline, both lasting for less than three months. The normal GFR rate is roughly 90-120 ml/min, though this varies according to body size, gender and muscle mass.

One hundred and thirty patients (5.2%) had a total of 144 episodes of ARF as defined by the study.

ARF was much more common in the first three months on HIV care after diagnosis. The incidence rate was 19.3% per year in the first three months (‘early onset’ ARF), 2.4% from month three to month six, 1.6% in the following six months, 1.25% in years two and three, and 0.5% after year three (‘late onset’ ARF). The incidence of early-onset ARF did not decline during the study. A total of 73 patients had early-onset ARF (3.2%) and 57 had late-onset ARF (2.5%).

Early-onset failure disproportionately affected black and heterosexual patients, those with a low CD4 count and lowest-ever (nadir) CD4 count, and patients with AIDS-defining conditions. Late-onset ARF affected more gay men and injecting drug users, and was more often caused by non-opportunistic bacterial infections (or the antibiotics used to treat them), including infections caused by injecting drug use and liver disease.

Patients with early-onset ARF were, after multivariate analysis which controlled for HIV risk group, ethnicity, CD4 count and AIDS status, three times more likely than patients without ARF to have had a CD4 nadir between 100 and 200 cells/mm³, 6.75 times more likely to have had a CD4 nadir under 100, and 6.72 times more likely to have had an AIDS diagnosis. Heterosexual patients were nearly 3.7 times more likely than gay men to develop ARF. They were also twice as likely to develop it as injecting drug users, and black patients 56% more likely than white patients, though in neither case were these differences statistically significant.

Late-onset ARF was, in contrast, less common among heterosexuals. Injecting drug users were 4.77 times more likely to develop late-onset ARF than heterosexuals and gay men 1.8 times more likely. Patients with hepatitis C were 3.36 times more likely to develop it. CD4 nadir and AIDS diagnosis remained as risks, with patients with a CD4 nadir below 100 cells/mm³ 5.82 times more likely to develop late-onset ARF and patients with an AIDS diagnosis 2.25 times more likely.

Causation appeared to vary according to whether ARF was early or late onset. Fifty-five per cent of patients with early-onset ARF had an AIDS defining infection compared with 14% with late-onset ARF, whereas 36% with early-onset ARF had a non-AIDS-defining infection compared with 48% of patients with late-onset ARF.

Liver disease was more common in patients with late-onset failure, and cancer (both AIDS- and non-AIDS-defining) equally common in both groups. A high proportion (73%) of patients with ARF, in either group, had taken nephrotoxic medication.

The most common infections in patients with early-onset ARF were the AIDS-defining conditions toxoplasmosis, PCP, and cryptococcal meningitis. In contrast the most common ones in late-onset patients were bacterial infections with Streptococcus and Staphylococcus aureus, manifesting as pneumonia, cellulitis, endocarditis, osteomyelitis and septicaemia.

Drugs used to treat these infections that are known to be nephrotoxic include amphotericin B, aminoglycosides, sulfadiazine, high-dose cotrimoxazole (Septrin), cidofovir, ganciclovir, flucloxacillin, vancomycin and rifampicin. High-dose non-steroidal anti-inflammatory drugs (NSAIDs) such as ibuprofen can also be nephrotoxic, as are a number of anti-cancer drugs.

HIV drugs were implicated in ten cases (14%) of late-onset ARF. These included six cases of indinavir-related kidney stones (all in the early part of the study period), two cases of lactic acidosis related to d4T-plus-ddI use, and two cases of ARF attributed to tenofovir. Despite the fact that 21% of all patients in the study took tenofovir, only 5.4% of patients experienced ARF while taking tenofovir.

However 21 out of 73 patients with early-onset ARF (29%) experienced it within 90 days of starting HAART. In six patients, IRIS may have contributed to ARF. Diseases implicated included two cases each of MAI and Cryptococcus and one each of PCP and reactivated hepatitis B. All the other patients also took potentially nephrotoxic drugs.

The study does not give overall mortality rates in patients nor estimates of the contribution of ARF to deaths and morbidity. However 31% of patients with early-onset ARF and 37% of patients with late-onset ARF died within 90 days of the ARF episode.

A quarter of patients with both early and late ARF experienced GFRs under 17-18 ml/min, where a rate of under 15 is considered to define kidney failure requiring dialysis or transplantation, and 29% of patients with early-onset ARF and 17% with late-onset in fact required dialysis. However the average duration of ARF was only 6-8 days. Dehydration was implicated in about 70% of cases and simple rehydration may have helped to reverse many cases.

In short, ARF was an under-documented cause of significant morbidity and possible mortality in a diverse population of patients with HIV at a London hospital and was especially common in newly-diagnosed patients, especially those who are diagnosed with AIDS. The authors comment that “in hospitalised patients, dehydration and use of NSAIDs should be avoided, and nephrotoxic medications should be used with caution, with close monitoring of renal function.”