| |
Therapy-associated severe hyperlactatemia rare in HIV patients
|
| |
| |
Study authors reported (see study report following Reuters report): ..... liver dysfunction was a predictor for severe hyperlactatemia....Among treatment-naive HIV-infected individuals who remained untreated during the observation period, HCV-coinfected individuals developed hyperlactatemia significantly more frequently than did HCV-seronegative persons....
NEW YORK (Reuters Health) - Although hyperlactatemia is often seen in HIV patients receiving potent combination antiretroviral therapy (ART) severe episodes are not common, Swiss researchers report in the September 1st issue of Clinical Infectious Diseases.
As senior investigator Dr. Rainer Weber told Reuters Health, "asymptomatic hyperlactatemia was frequently observed among persons on ART, but severe hyperlactatemia or life-threatening lactic acidosis were rarely found."
To investigate the prevalence of the disorder, Dr. Weber of University Hospital Zurich and colleagues followed patients already receiving ART, treatment-naive patients beginning ART and patients not receiving ART.
In 4788 person-years of follow-up, more than 22,000 lactate assessments were performed in 1566 patients. In total 662 (42.3%) patients had at least one lactate measurement of greater than 2.4 mmol/L and 49 (3.8%) had severe hyperlactatemia as indicated by levels greater than 5 mmol/L. Severe hyperlactatemia was seen only in patients receiving ART.
The incidence of hyperlactatemia was 227 per 1000 person years in patients receiving ART and 59 per 1000 person years in those not receiving ART. Because of changing prescription patterns, the overall incidence dropped from 459 cases to 85 cases per 1000 patient-years of follow-up.
Significant risk factors for severe hyperlactatemia were regimens containing stavudine and didanosine (hazard ratio 6.65) or efavirenz (hazard ratio, 2.85). Non-fatal lactic acidosis was seen in four patients all of whom were being treated with regimens containing stavudine and didanosine.
The incidence of hyperlactatemia decreased substantially after stavudine-based regimens were withdrawn, the investigators point out, but these findings may remain important because the increasing use of generic formulations of stavudine in developing countries.
Given the rarity of severe hyperlactatemia, the researchers conclude that monitoring is indicated primarily "for persons receiving stavudine and didanosine and for persons who are symptomatic."
Clin Infect Dis 2005;41:721-728.
Risk Factors for and Outcome of Hyperlactatemia in HIV-Infected Persons: Is There a Need for Routine Lactate Monitoring?
Clinical Infectious Diseases Sept 1 2005;41:721-728
Alexander Imhof,1 Bruno Ledergerber,1 Huldrych F. Günthard,1 Stefan Haupts,1 Rainer Weber,1 and the Swiss HIV Cohort Studya
1Division of Infectious Diseases and Hospital Epidemiology, University Hospital Zürich, Switzerland
ABSTRACT
Background. Lactic acidosis is a rare but life-threatening complication of combination antiretroviral therapy (CART). Asymptomatic or mildly symptomatic episodes of hyperlactatemia are more frequent, but their clinical relevance is unknown.
Methods. The incidences of, risk factors for, and courses of hyperlactatemia and lactic acidosis were prospectively assessed in the following 3 groups at the Zurich center of the Swiss HIV Cohort Study: persons already receiving CART at baseline, treatment-naive persons who initiated CART during the observation period, and persons who received no CART before or during the observation period.
Results. During 4788 person-years of follow-up, a total of 22,678 lactate assessments were performed for 1566 persons; 662 (42.3%) had at least 1 lactate level measurement of >2.4 mmol/L, and 49 (3.1%) had severe hyperlactatemia (lactate level of >5.0 mmol/L). The incidence of hyperlactatemia was 227 cases (95% confidence interval [CI], 210-245) and 59 cases (95% CI, 38-93) per 1000 person-years of follow-up among persons with and persons without CART, respectively. During the observation period, the incidence decreased from 459 cases (95% CI, 415-508) to 85 cases (95% CI, 76-107) per 1000 person-years of follow-up, respectively, because of changing CART prescription patterns. Severe hyperlactatemia occurred in treated persons only. In multivariable Cox proportional hazards models, significant risk factors for severe hyperlactatemia were regimens containing stavudine and didanosine (hazard ratio [HR], 6.65; 95% CI, 2.70-16.3) and regimens containing efavirenz (HR, 2.85; 95% CI, 1.31-6.21). Lactic acidosis was diagnosed in 4 of 1566 persons, all of whom were receiving stavudine and didanosine.
Conclusions. Hyperlactatemia was frequently observed in all 3 groups, but severe hyperlactatemia and lactic acidosis were rarely observed among persons who received CART. Lactate monitoring appears to be indicated primarily for persons receiving stavudine and didanosine and for persons who are symptomatic. Long-term follow-up is needed to investigate the risk of novel treatment regimens for hyperlactatemia.
AUTHOR DISCUSSION
We analyzed 22,678 plasma lactate assessments for 1566 HIV-infected persons who were prospectively followed-up in the SHCS for 4788 person-years. A total of 42.3% of all persons had at least 1 elevated lactate level, 18.5% had elevated lactate levels (>2.4 mmol/L) on at least 2 consecutive occasions, and 3.1% developed severe hyperlactatemia (lactate level of >5 mmol/L). The incidence of hyperlactatemia among persons receiving CART was significantly higher than that for untreated individuals but decreased substantially during the observation period, because of changing prescription patterns of CART regimens. Four persons developed nonfatal, CART-associated lactic acidosis; regimens for all 4 contained stavudine and didanosine. Although the incidence of hyperlactatemia decreased substantially in our cohort after stavudine-based regimens were withdrawn, our findings may remain important, because of increasing use of the low-cost generic formulation of stavudine for the treatment of HIV-infected persons in developing countries.
In our cohort, predictors for severe hyperlactatemia included age, severe immunodeficiency, liver dysfunction, and drug regimens, particularly those containing didanosine and/or stavudine and, unexpectedly, those containing efavirenz. Furthermore, hyperlactatemia was also associated with increased waist-hip ratio (i.e., fat accumulation). Among treatment-naive HIV-infected individuals who remained untreated during the observation period, HCV-coinfected individuals developed hyperlactatemia significantly more frequently than did HCV-seronegative persons.
The strengths of this observational database are the large sample size, the duration of observation period, and inclusion of a large proportion of women and persons from all groups at risk for HIV infection. Furthermore, persons enrolled in cohort studies tend to be more representative of the community of HIV-infected people and available for longer durations of follow-up than do persons enrolled in clinical trials. A limitation of our study is that nonspecific signs and symptoms, which are not related to a specific complication of HIV infection, were not systematically and prospectively collected. Creatinine clearance (which has been described as a predictor for lactic acidosis [19]) was not assessed, but, on the basis of results of routine monitoring of creatinine levels, renal insufficiency is very rare in our mainly white population. Finally, we cannot exclude the possibility that our results were distorted by residual confounding or by unmeasured confounding factors.
What are the new findings of our cohort study? First, our estimates of the incidence of hyperlactatemia are higher than data reported elsewhere [15, 16, 20], but the estimates decreased substantially during the observation period, because of changing NRTI prescription patterns. An investigation involving 4361 lactate measurements in a cohort of 2069 HIV-infected persons, 1239 of whom were receiving therapy, reported that 8.7% of individuals had lactate levels of >2.4 mmol/L and that 0.4% had severe hyperlactatemia [15]. A previous SHCS study of the prevalence of hyperlactatemia involving 880 persons found that 8.3% of individuals had elevated lactate levels of >2.0 mmol/L. Similar to our results, a significant association of hyperlactatemia with didanosine alone and with didanosine plus stavudine was documented, but stavudine was not an independent predictor in a multivariable model [21]. Associations between hyperlactatemia and use of stavudine [9, 16, 22], didanosine [15], and both [22] were also shown in other studies. Elsewhere, an association between hyperlactatemia and ritonavir similar to that found in our study was reported for persons treated with ritonavir in a boosted PI-containing regimen and for persons treated with double PI-containing regimens [16].
Second, we studied the possible association between efavirenz and hyperlactatemia. Previously, an increased but not significant OR (OR, 1.6; 95% CI, 0.7-3.6 [16]) and an increased hazard ratio of hyperlactatemia (HR, 1.45; 95% CI, 0.97-2.16 [15]) were reported for efavirenz. Our multivariable Cox model including different drug classes revealed an association between this nonnucleoside reverse-transcriptase inhibitor and hyperlactatemia, but in another model investigating frequently used CART regimens, the effect of efavirenz in a 2-class combination with 2 NRTIs (excluding didanosine and/or stavudine) was not conclusive, although point estimates (HR, 1.34; 95% CI, 0.80-2.27) remained comparable to the values of the first model (HR, 1.46; 95% CI, 1.15-1.87). The specific mechanism of this association is unknown. In vitro data using a hepatoma cell line did not find that efavirenz affected lactate output from their cells, whereas various nucleoside analogues had dramatic effects in this regard, particularly when used in combination [13]. A recently suggested mechanism that is based on in vitro work may involve suppression of the lipogenic pathway by efavirenz [23].
Third, in addition to antiretroviral drugs, demographic and clinical variables, including age [22], lower CD4 lymphocyte counts [19], and elevated liver enzyme levels (which is a surrogate for liver dysfunction and consecutive altered lactate clearance) were also independent predictors for severe hyperlactatemia. Furthermore, an increased waist-hip ratio, a surrogate for lipodystrophy, showed a significant association with hyperlactatemia. The association between NRTI-associated hyperlactatemia and lipodystrophy has led to the hypothesis that mitochondria may play a key role in the pathogenesis of body fat changes, possibly through release of apoptosis mediators, which, in turn, lead to peripheral fat loss [24]. However, conflicting clinical findings have been published on this issue; an association between hyperlactatemia and lipodystrophy has been described [4, 25], whereas other studies were not able to demonstrate such correlations [26].
We demonstrated that increased lactate levels are a common but often transient phenomenon in persons receiving a consistent regimen of CART. However, symptomatic hyperlactatemia can have a negative effect on quality of life, adherence to therapy, and, ultimately, the ability to receive effective long-term CART. Evidence is mounting that hyperlactatemia and lactic acidosis, in the presence of hepatic steatosis or hepatitis, are associated with increased morbidity and mortality [4, 6, 18]. Serum transaminase levels can remain nearly normal, even in cases of severe liver steatosis, mitochondrial liver toxicity, and lactic acidosis. However, we found elevated liver enzyme levels to be predictive of mild or severe hyperlactatemia.
On the basis of our findings, one might suggest that routine lactate assessment is too expensive and that is not sufficiently predictive for severe adverse events, including lactic acidosis. In routine practice, it may be sufficient to monitor plasma lactate levels only in persons at risk, including elderly persons, persons with altered liver function, persons receiving didanosine or stavudine, and persons receiving concomitant therapy for HIV and HCV coinfection. Still, life-threatening lactic acidosis, although rare, can be prevented by lactate monitoring. Moreover, there is a need for continued long-term prospective studies for many reasons, including the following: long-term adverse effects of novel CART regimens are unknown; concomitant treatment of coinfections or comorbidities are emerging, frequently involving novel combinations of multiple drugs with unknown interactions; possible toxicities; and possible risk for hyperlactatemia.
RESULTS
Study population. A total of 1566 persons were included in the present analysis, of whom 1178 were already receiving CART as of 1 August 1999. A total of 214 persons were treatment naive as of 1 August 1999 and started CART afterward, and 174 individuals never received CART.
Antiretroviral drug prescription pattern, and prevalence of hyperlactatemia. The prescription pattern of antiretroviral drugs changed between 1999 and 2004. The proportion of persons receiving regimens containing stavudine and/or didanosine decreased significantly, whereas the proportion who received efavirenz increased. These changes in prescription patterns were paralleled by a pronounced reduction in the incidence of hyperlactatemia.
During 4788 person-years of follow-up, 22,678 lactate measurements were available for 1566 persons, of whom 662 (42.3%) had lactate levels of >2.4 mmol/L, and 49 (3.1%) had at least 1 episode of severe hyperlactatemia (i.e., lactate levels of >5 mmol/L). Lactate levels of >5 mmol/L were all observed in persons who received CART. Most treated persons (96.8%) received ⩾3 antiretroviral drugs. The median value of the first lactate measurement was 1 mmol/L (interquartile range [IQR], 0.8-1.4) for persons who received no treatment before or during the study and 1.3 mmol/L (IQR, 1.0-1.8) for those receiving treatment at baseline.
Course of hyperlactatemia. The positive predictive value for having an elevated lactate level on 2 consecutive occasions was low (43.8% of all persons, 44.8% of those who received CART, and 21.1% of those who did not receive CART). The negative predictive value for having a low lactate level on 2 consecutive occasions was high (93.68% of all persons, 93.3% of those who received CART, and 97.7% of those who did not receive CART). An initial value of >2.4 mmol/L was often followed by a value within the normal range. Only 290 (18.5%) of 1566 persons, 284 (98%) of whom were receiving CART, had at least 2 consecutive elevated lactate levels.
Incidence of hyperlactatemia. During the observation period, 19 (11%) of the 174 persons who never were treated, 72 (35%) of 207 persons who initiated CART during the study (7 persons with elevated lactate levels before treatment initiation were excluded), and 564 (48%) of the 1178 persons who were already receiving CART at baseline developed hyperlactatemia, with incidences of 59 (95% CI, 38-93), 229 (95% CI, 182-288), and 226 (95% CI, 209-246) cases, respectively, per 1000 person-years of follow-up. Severe hyperlactatemia occurred in treated persons only, with an incidence of 9.5 cases per 1000 person-years of follow-up (95% CI, 7.0-12.9).
The incidences of hyperlactatemia, at least 2 consecutive values >2.4 mmol/L, and severe hyperlactatemia decreased significantly during the observation period, as shown in figure 1. The incidence of having at least 2 consecutive elevated lactate levels decreased from 172 cases per 1000 person-years of follow-up (95% CI, 147-200) in the first observation period (August 1999 through December 2000) to 22 cases per 1000 person-years of follow-up (95% CI, 15-33) in the last period (January 2003 through February 2004).
Risks for hyperlactatemia. Table 2 shows the results of univariable and multivariable Cox models of the risk for developing hyperlactatemia during the observation period among persons already receiving CART at baseline and those who initiated CART during the study. The risks for developing hyperlactatemia were older age, increased waist-hip ratio, high CD4 cell count, liver dysfunction, and receipt of various CART regimens. A decreased risk for hyperlactatemia during the observation period was observed.
Most drug regimens associated with hyperlactatemia included stavudine. Independent predictors for hyperlactatemia included efavirenz, boosted PI regimens, and double PI-containing regimens.
In addition to the analyses reported in table 2, we compared the risk of hyperlactatemia associated with different frequently used CART regimens in a multivariable Cox model (which also included sex, age, waist-hip ratio, CD4 cell count, liver function status, and HCV infection status). The effect of efavirenz in a 2-class regimen with 2 NRTIs (excluding didanosine and stavudine) was not conclusive, although point estimates (HR, 1.34; 95% CI, 0.80-2.27) remained comparable to the values in table 2 (HR, 1.46; 95% CI, 1.15-1.87). This may be partly due to smaller sample size, but we cannot exclude residual confounding by stavudine and/or didanosine. The variable "3-class regimen without stavudine and didanosine" remained a significant risk (HR, 1.96; 95% CI, 1.18-3.26). All other frequently used CART regimens (excluding those containing didanosine and stavudine) were neither associated with hyperlactatemia nor with a lower risk for elevated plasma lactate levels in this multivariable Cox model.
Thus, the probability for developing hyperlactatemia was clearly associated with antiretroviral therapy, as also depicted in the Kaplan-Meier plots. However, persons who didn't receive antiretroviral therapy also developed hyperlactatemia over time, particularly those with chronic HCV infection.
Severe hyperlactatemia and lactic acidosis. Of 41 individuals who had severe hyperlactatemia, 4 developed lactic acidosis, but none died. The 4 persons with acidosis had been receiving CART for a mean of 9.7 months (range, 2.8-16.1 months). Symptoms at the time of presentation included lethargy, nausea, vomiting, and abdominal pain. All persons were receiving stavudine and didanosine, as well as a protease inhibitor (PI) and/or a non-NRTI. Two persons underwent lactate level assessment before onset of acidosis. For both, previous samples obtained during a 6-month period were >2.4 mmol/L (range, 2.4-5.8 mmol/L in the first person and 3.8-8.7 mmol/L in the second person). The first person was mildly symptomatic during the week before developing acidosis. The second had flulike symptoms during acidosis but was asymptomatic before acidosis onset. No previous lactate level assessments were available in the other 2 acidotic persons, who presented with lactate levels of 5.8 mmol/L and 7.4 mmol/L.
A total of 37 persons presented with severe hyperlactatemia (range, 5.1-7.7 mmol/L) without acidosis. In 11 persons, hyperlactatemia was a consequence of the following complications: acute bacterial infections with septicemia (7 persons), liver failure due to hepatocellular carcinoma or chronic HCV infection (3 persons), and respiratory and cardiac arrest (1 injection drug user, in whom lactic acidosis was instead the consequence rather than the cause of resuscitation). Twenty-six persons were asymptomatic at the time of severe hyperlactatemia. Of these, 17 had lactate levels >5 mmol/L only once. Nineteen persons (51.4%) were receiving regimens containing stavudine and didanosine.
Mortality. A total of 93 (5.9%) of 1566 persons died. Twelve persons died of AIDS-defining infection, 9 died of AIDS-defining malignancy, 12 died of non-AIDS-defining infection, 5 died of myocardial infarction, 6 died of other cardiovascular diseases, 4 died of drug intoxication, 18 died of liver disease, 12 died of non-AIDS-defining cancer, 6 died of suicide, 1 died of an accident, and 8 died of unknown causes. Of the 93 persons who died, 48 (52%) had normal lactate levels within 3 months before death; 20 (22%) had hyperlactatemia; and 25 (27%) did not undergo lactate assessment within 3 months before death.
Within 3 weeks before death, 29 persons had normal lactate levels, 49 did not undergo lactate assessment, and 15 had levels >2.4 mmol/L. Of the latter 15 persons, 9 deaths were attributable to liver diseases, 2 were attributable to sepsis, 1 was due to endocarditis, 1 was due to respiratory arrest (associated with injection drug use), and 2 were due to unknown causes.
Costs of lactate monitoring. Collection of a specimen for lactate analysis and processing the specimen in a fluoride tube costs approximately US$20. A total of 22,678 samples were analyzed to diagnose 41 episodes of severe hyperlactatemia. The total cost of these lactate analyses was US$462,007. Therefore, US$11,268 was spent to detect each severe hyperlactatemic episode.
|
|
| |
| |
|
|
|
