Compliance with national guidelines for HIV treatment and its association with mortality and treatment outcome: a study in a Spanish cohort

Authors


  • Data from this study were presented at the 11th International Congress on Drug Therapy in HIV Infection, November 2012, Glasgow, UK (P267) and published as an abstract in the Journal of the International AIDS Society 2012, 15 (Suppl 4): 18243.

Abstract

Objectives

The aim of the study was to assess the adequacy of initial antiretroviral therapy (ART), in terms of its timing and the choice of regimens, according to the Spanish national treatment guidelines [Spanish AIDS Study Group−National Plan for AIDS (GeSIDA-PNS) Guidelines] for treatment-naïve HIV-infected patients.

Methods

A prospective cohort study of HIV-positive ART-naïve subjects attending 27 centres in Spain from 2004 to 2010 was carried out. Regimens were classified as recommended, alternative or nonrecommended according to the guidelines. Delayed start of treatment was defined as starting treatment later than 12 months after the patient had fulfilled the treatment criteria. Multivariate logistic and Cox regression analyses were performed.

Results

A total of 6225 ART-naïve patients were included in the study. Of 4516 patients who started treatment, 91.5% started with a recommended or alternative treatment. The use of a nonrecommended treatment was associated with a CD4 count > 500 cells/μL [odds ratio (OR) 2.03; 95% confidence interval (CI) 1.14–3.59], hepatitis B (OR 2.23; 95% CI 1.50–3.33), treatment in a hospital with < 500 beds, and starting treatment in the years 2004–2006. Fourteen per cent of the patients had a delayed initiation of treatment. Delayed initiation of treatment was more likely in injecting drug users, patients with hepatitis C, patients with higher CD4 counts and during the years 2004–2006, and it was less likely in patients with viral loads > 5 log HIV-1 RNA copies/ml. The use of a nonrecommended regimen was significantly associated with mortality [hazard ratio (HR) 1.61; 95% CI 1.03–2.52; P = 0.035] and lack of virological response.

Conclusions

Compliance with the recommendations of Spanish national guidelines was high with respect to the timing and choice of initial ART. The use of nonrecommended regimens was associated with a lack of virological response and higher mortality.

Introduction

Clinical guidelines are available for an increasing array of diseases [1, 2]. Despite this, physicians' adherence to clinical guidelines is highly variable and is influenced by multiple factors, such as guideline complexity, implementation strategies, physician characteristics, general workload, institutional support, and the patient's comorbidities [1].

Guidelines for the treatment of HIV infection are available in many countries [3-5]; however, there is little evidence assessing compliance with clinical practice guidelines for antiretroviral treatment and its impact on clinical outcomes. Among the few studies that have investigated this issue [6-17], only two [7, 13] evaluated the relationship of adherence to treatment guidelines with immunovirological outcome. Most importantly, none of these studies has evaluated its relationship with mortality.

The Spanish health care system provides free-of-charge health care to all HIV-infected patients. The Spanish national guidelines for antiretroviral treatment are published by the Spanish AIDS Study Group (GeSIDA) and the National Plan for AIDS (PNS): they provide guidance on antiretroviral treatment of HIV-infected adults, and give recommendations on when to start treatment in naïve patients, as well as recommended antiretroviral regimens [18].

The aim of this study was to assess compliance with national guidelines for the treatment of naïve patients in a multicentre Spanish cohort [the Cohort of the Spanish AIDS Research Network (CoRIS)], and to investigate whether adherence to the guidelines is associated with better outcomes. The specific aims were to evaluate the proportion of patients treated according to the guidelines' recommendations in terms of initial antiretroviral regimens and appropriate timing of treatment initiation, to investigate factors associated with the prescription of a nonrecommended treatment and with a delayed initiation of treatment, and to assess the prognosis of patients receiving nonrecommended treatments in terms of mortality and of virological and immunological response.

Methods

Patients and data collection

Patients were selected from CoRIS, which has been described in previous publications [19, 20]. CoRIS is a prospective multicentre cohort of HIV-positive treatment-naïve subjects aged > 13 years, recruited from 28 centres in the Spanish public health care system. Patients were included from January 2004 to October 2010, which was the date of administrative censoring for this study.

We collected data for the following variables: age at enrolment and at treatment initiation, sex, transmission category [heterosexual contact, men who have sex with men (MSM), injecting drug user (IDU) and other/unknown], country of origin, education level, CD4 cell count and viral load at enrolment and at treatment initiation, hepatitis B at enrolment [defined as positive hepatitis B virus (HBV) surface antigen (HbsAg)], hepatitis C at enrolment [defined as positive hepatitis C virus (HCV) antibodies], treatment administered, date of treatment initiation (2004–2006, 2007–2008 and 2009–2010), date of stage B or C clinical events according to the Centers for Disease Control and Prevention (CDC) classification [21], hospital of follow-up, date of last clinical follow-up visit or death, and cause of death.

Additionally, we collected information on primary antiretroviral resistance (before treatment initiation) for a subset of patients in which this information was available. This variable was registered as the presence or absence of resistance to any antiretroviral group (nucleoside or nonnucleoside reverse transcriptase inhibitors or protease inhibitors). Twenty-three centres in CoRIS contribute naïve FASTA sequences from routine resistance testing. Resistance-associated mutations were evaluated following the World Health Organization (WHO) surveillance list [22]; the detailed methodology has been described elsewhere [23, 24].

During the study period, the guidelines were updated several times, and the changes introduced were taken into account from the day of publication for each period [25-32]. Treatments administered were classified as recommended, alternative or not recommended, according to the guidelines' ‘what to start with’ recommendations. The recommendations for ‘recommended’ and ‘alternative’ drug combinations for treatment initiation in each period are shown in Table 1a. Any treatment that was not classified as recommended or alternative was considered nonrecommended. We analysed only the first treatment administered; any further treatment changes were ignored.

Table 1. Recommendations of the Spanish guidelines. (a) Recommended and alternative drug combinations for treatment initiation in naïve HIV-infected patients, updated during the study period; (b) indications for initiating antiretroviral therapy for patients with chronic HIV infection according to CD4 cell count
(a)
January 2004 to October 2004*
Recommendedd4T + ddI, d4T + 3TC, ddI + 3TC, ZDV + ddI, ZDV + 3TC, ZDV + ddCABC, EFV, IDV, NFV, NVP, RTV, APV/r, IDV/r, LPV/r, SQV/rOne drug/combination from the second column and one drug/combination from the third column
November 2004 to December 2005
RecommendedZDV, ABC, TDF, ddI3TC, FTCEFV, NVP, LPV/rOne drug from the second column, one drug from the third column and one drug from the fourth column
AlternativeATV/r + 2NRTIs, ATV + 2NRTIs, FOS/r + 3TC + ABC, FOS + 3TC + ABC, IDV/r + 2NRTIs, NFV + 2NRTIs, SQV/r + 2NRTIs, ABC + 3TC + ZDV
January 2006 to December 2006
RecommendedZDV, ABC, TDF, ddI3TC, FTCEFV, NVP, LPV/rOne drug from the second column, one drug from the third column and one drug from the fourth column
AlternativeATV/r + 2NRTIs, FOS/r + 3TC + ABC, IDV/r + 2NRTIs, NFV + 2NRTIs, SQV/r + 2NRTIs, ABC + 3TC + ZDV
January 2007 to January 2008
RecommendedTDF, ABC, ZDV3TC, FTCEFV, LPV/r, FPV/rOne drug from the second column, one drug from the third column and one drug from the fourth column
AlternativeddI, d4TNVP, ATV/r, SQV/r, ATV, NFV
AlternativeABC + 3TC + ZDV
February 2008 to September 2008
RecommendedTDF, ABC3TC, FTCEFV, LPV/r, FPV/r, SQV/rOne drug from the second column, one drug from the third column and one drug from the fourth column
AlternativeZDV, ddI, d4TNVP, ATV/r, LPV/r, FPV/r, ATV, FPV
AlternativeABC + 3TC + ZDV, ABC + 3TC + ZDV + TDF, maraviroc + 3TC + ZDV
October 2008 to January 2009
RecommendedTDF, ABC3TC, FTCEFV, LPV/r, FPV/r, SQV/r, ATV/rOne drug from the second column, one drug from the third column and one drug from the fourth column
AlternativeZDV, ddI, d4TNVP, LPV/r, FPV/r, ATV, FPV
AlternativeABC + 3TC + ZDV, ABC + 3TC + ZDV + TDF, maraviroc + 3TC + ZDV
February 2009 to December 2009
RecommendedTDF, ABC3TC, FTCEFV, LPV/r, FPV/r, SQV/r, ATV/r, DRV/rOne drug from the second column, one drug from the third column and one drug from the fourth column
AlternativeZDV, ddI, d4TNVP, LPV/r, FPV/r, ATV, FPV
AlternativeABC + 3TC + ZDV, ABC + 3TC + ZDV + TDF, maraviroc + 3TC + ZDV
January 2010
RecommendedTDF, ABC3TC, FTCEFV, NVP, LPV/r, FPV/r, SQV/r, ATV/r, DRV/r, raltegravirOne drug from the second column, one drug from the third column and one drug from the fourth column
AlternativeZDV, ddIATV, FPV, FPV/r, maraviroc
AlternativeABC + 3TC + ZDV, ABC + 3TC + ZDV + TDF
(b)
PeriodCD4 count*
< 200 cells/μL< 350 cells/μL350–500 cells/μL> 500 cells/μL
  1. During the period from January 2010 to December 2010, the guidelines recommended considering treatment initiation in patients with CD4 count > 500 cells/μL in the case of having a serodiscordant partner.
  2. *Initiation of antiretroviral therapy is always indicated in the presence of symptoms of Centers for Disease Control and Prevention (CDC) stage B or C events, regardless of CD4 count.
Feb 2004 to Jan 2008Always recommendedRecommended on most occasionsNot recommendedNot recommended
Feb 2008 to Jan 2009Consider in the presence of: viral load > 100 000 copies/ml, CD4 percentage > 14%, cirrhosis, hepatitis B when its treatment is indicated, or chronic hepatitis CNot recommended
Feb 2009 to Oct 2010Always recommendedRecommended in the presence of: viral load > 100 000 copies/ml, CD4 percentage > 14%, cirrhosis, hepatitis B when its treatment is indicated, chronic hepatitis C, age > 55 years, high cardiovascular risk, or HIV nephropathyConsider in the presence of: viral load 100 000 copies/ml, CD4 percentage > 14%, cirrhosis, hepatitis B when its treatment is indicated, chronic hepatitis C, age > 55 years, high cardiovascular risk, or HIV nephropathy

The guidelines' recommendations for initiating antiretroviral treatment (‘when to start’) are shown in Table 1b. The date of treatment indication was defined as the day on which a patient first fulfilled the guidelines' criteria to recommend treatment initiation. We considered the criteria for which the guidelines explicitly said that treatment was ‘recommended’, but not those for which treatment was to be ‘considered’. Data on HIV nephropathy, cardiovascular risk or cirrhosis are not available in our cohort and thus were not used as criteria for treatment initiation. Delayed treatment initiation was defined as initiating treatment more than 1 year after the date when treatment was indicated by the guidelines, or not initiating treatment at all when it was indicated by the guidelines. Patients were included in this analysis only if they had at least 12 months of follow-up time after treatment was indicated.

Virological and immunological responses were defined as undetectable viral load and a CD4 count increase of at least 100 cells/μl, respectively, after 1 year of treatment (closest value between 10 and 14 months). Undetectable viral load was defined as a viral load of < 50 HIV-1 RNA copies/ml. Patients who died during the first year of treatment were assumed to have no virological or immunological response.

Statistical analysis

Baseline characteristics of patients receiving nonrecommended treatments were compared with those of all other patients (receiving recommended or alternative treatments) using χ2 or Fisher's exact tests.

Predictors of receiving a nonrecommended treatment and of delayed start of treatment were assessed using a multivariate logistic regression model. We used a backward stepwise selection method, and variables with a P-value > 0.10 were excluded from the model. Patients with primary HIV infection were excluded from the analysis of delayed start of treatment.

Virological and immunological outcomes were compared in patients receiving nonrecommended treatments and those receiving a recommended/alternative treatment using a multivariate logistic regression model. Mortality after treatment initiation was compared in patients receiving nonrecommended treatments and those receiving a recommended/alternative treatment with multivariate Cox regression. Time was calculated from the date of start of treatment to the date of death or date of last clinical follow-up.

Robust methods were used to estimate confidence intervals (CIs), assuming correlation between the subjects in each centre. Wald tests were used to calculate P-values. All analyses were performed with a 95% confidence level. Statistical analyses were performed in stata version 9.0 (StataCorp, College Station, TX).

Ethics

All patients signed informed consent forms. The programme was approved by the Institutional Review Boards of the participating centres [19, 20].

Results

Patients and baseline characteristics

A total of 6225 treatment-naïve patients from 28 hospitals were included in CoRIS between the years 2004 and 2010, of whom 4516 patients started antiretroviral treatment. The baseline characteristics of the study population are shown in Table 2.

Table 2. Baseline characteristics of the study population
  1. Values are n (%) unless otherwise stated.
  2. IQR, interquartile range.
Total n6225
Female gender1336 (21.5)
Age at enrolment (years) [median (IQR)]35 (29−43)
Age at enrolment 
18–30 years1185 (30.3)
31–40 years2405 (38.6)
41–50 years1342 (21.6)
>50 years593 (9.5)
Route of transmission 
Heterosexual2266 (36.4)
Men who have sex with men2809 (45.1)
Injecting drug user907 (14.5)
Other/unknown243 (3.9)
CD4 count at enrolment (cells/μl) [median (IQR)]333 (160–529)
CD4 count at enrolment 
≤ 50 cells/μL693 (11.1)
51–200 cells/μL1193 (19.2)
201–350 cells/μL1327 (21.3)
351–500 cells/μL1226 (19.7)
> 500 cells/μL1719 (27.6)
Unknown67 (1.8)
Viral load at enrolment (log10 copies/mL) [median (IQR)]4.6 (4–5.2)
Viral load at enrolment 
< 4.0 log10 copies/mL1526 (24.5)
4.0–5.0 log10 copies/mL2513 (40.4)
> 5.0 log10 copies/mL2107 (33.8)
Unknown79 (1.3)
Chronic hepatitis C 
No4719 (75.8)
Yes1104 (17.7)
Unknown402 (6.5)
Chronic hepatitis B 
No5508 (88.5)
Yes304 (4.9)
Unknown413 (6.6)
Education 
University/secondary education2833 (45.5)
Primary/no education2293 (36.8)
Unknown1099 (17.7)
Country of origin 
Spain4349 (69.9)
Country other than Spain1876 (30.1)
Number of hospital beds 
>10003075 (49.4)
500–10002186 (35.1)
<500964 (15.5)

Information on primary resistance to antiretroviral drugs was available for 1068 patients, of whom 81 (7.6%) had resistance to at least one class of drug: 19 patients (1.8%) had resistance to protease inhibitors, 41 (3.8%) had resistance to nucleoside reverse transcriptase inhibitors and 39 (3.6%) had resistance to nonnucleoside reverse transcriptase inhibitors.

Analysis of factors associated with the prescription of nonrecommended treatments

During the study period, 4516 patients started treatment. Of these patients, 3592 (79.5%), 540 (12%) and 384 (8.5%) started with a recommended, alternative and nonrecommended treatment, respectively. The number and proportion of patients starting treatment with a recommended, alternative and nonrecommended regimen each year are shown in Figure 1.

Figure 1.

Proportion of patients starting treatment with a recommended, alternative or not recommended regimen each year.

Predictors of receiving a nonrecommended treatment were assessed using a multivariable logistic regression model. The use of a nonrecommended treatment was significantly associated with a CD4 count > 500 cells/μl at treatment initiation, HBV infection, treatment in a hospital with < 500 beds, and starting treatment in the years 2004 to 2006 (Table 3). After adjusting for other variables, there was no significant association of having a nonrecommended treatment with age, gender, route of transmission, HCV infection, country of origin, education, or viral load at treatment initiation.

Table 3. Predictors of receiving a nonrecommended treatment (n = 4516)
VariableNot recommended [n (%)/total]UnivariableMultivariable
OR (95% CI)POR (95% CI)P
  1. CI, confidence interval; OR, odds ratio.
CD4 count at treatment initiation  <0.001 < 0.001
≤ 50 cells/μL55 (9)/6101 1
51–200 cells/μL101 (8.6)/11800.94 (0.61–1.24) 0.96 (0.61–1.51)
201–350 cells/μL99 (6.6)/14920.71 (0.47–1.07) 0.83 (0.55–1.25)
351–500 cells/μL49 (9.8)/5021.09 (0.70–1.70) 1.29 (0.80–2.10)
>500 cells/μL33 (15.2)/2171.81 (1.01–3.25) 2.03 (1.14–3.60)
Unknown47 (9.1)/5151.01 (0.59–1.74) 1.08 (0.64–1.82)
Hepatitis B  <0.001 < 0.001
No326 (8.1)/40341 1
Yes41 (16.5)/2482.25 (1.59–3.20) 2.24 (1.50–3.33)
Unknown17 (7.3)/2340.89 (0.53–1.48) 0.95 (0.53–1.72)
Date of treatment initiation  <0.001 < 0.001
2009–201089 (6.2)/14291 1
2007–200862 (4)/15360.63 (0.40–1.00) 0.63 (0.41–0.97)
2004–2006233 (15)/15512.66 (1.74–4.07) 2.64 (1.72–4.07)
Number of hospital beds  0.154 0.043
>1000157 (7)/22361 1
500–1000140 (8.9)/15821.28 (0.67–2.47) 1.19 (0.63–2.22)
<50087 (12.5)/6981.89 (0.99–3.59) 1.70 (1.09–2.65)

As some clinicians might be aware of new updates to the guidelines before they are released, we repeated the analysis to consider the regimens that were started ahead of time, before they became recommended in the guidelines. Of the 384 patients who started treatment with a nonrecommended regimen, 106 (27.6%) started with a regimen that was nonrecommended at the time of its prescription, but which was included in the clinical guidelines as ‘recommended’ or ‘alternative’ in later periods. For this analysis, treatments administered at a time when they would be ‘nonrecommended’, but which later would become ‘recommended’ or ‘alternative’ (which we have called ‘later recommended’), were considered in the same group as the recommended/alternative regimens. We performed the multivariate logistic regression assessing predictors of starting a nonrecommended treatment compared with having a recommended, alternative or later recommended treatment: our results were not changed significantly and the variables associated with receiving a nonrecommended treatment were the same (data not shown).

In order to investigate whether the presence of primary resistance to antiretrovirals may have influenced treatment decisions, we performed a subanalysis with the 1068 patients who had this information available. Seven (13.7%) of 51 patients receiving nonrecommended treatments had primary resistance to at least one class of antiretroviral, compared with 74 (7.3%) of 1017 patients receiving recommended/alternative treatments (P = 0.100). In univariate analysis of predictors of prescription of nonrecommended treatments, the odds ratio (OR) for primary resistance was 2.03 (95% CI 0.62–6.64; P = 0.243). After adjusting for age, gender, route of transmission, country of origin, education, viral load, hepatitis B and C, date of treatment and number of hospital beds, the OR was 1.77 (95% CI 0.54–5.87; P = 0.384).

Analysis of factors associated with delayed start of treatment

Among the 6225 patients included in the study, 121 had primary HIV infection at the time of enrolment and were excluded. Of the remaining 6104 patients, 3530 patients fulfilled treatment criteria. A total of 484 patients were excluded because they had a follow-up time of < 12 months. Among the 3046 patients available for the analysis, 2620 (86%) had timely treatment (were treated during the first year after fulfilling treatment criteria as defined by the guidelines) and 426 (14%) had delayed treatment (were treated > 1 year after fulfilling treatment criteria as defined by the guidelines, or were not treated at all).

Factors associated with a delayed start of treatment are shown in Table 4. IDUs, patients with hepatitis C and patients fulfilling criteria for treatment in the years 2004–2006 had a higher risk of delayed start of treatment, whereas the risk was lower in patients with viral load > 5 log copies/ml at enrolment. The risk of delayed treatment increased with increasing CD4 count at enrolment. After adjusting for other risk factors, there was no association between treatment delay and sex, age, country of origin, education, HBV infection or number of hospital beds.

Table 4. Factors associated with delayed start of treatment (n = 3046)
VariableDelayed treatment [n (%)/total]UnivariableMultivariable
OR (95% CI)POR (95% CI)P
  1. CI, confidence interval; OR, odds ratio.
Route of transmission  <0.001 0.036
Heterosexual124 (10.4)/11891 1
Men who have sex with men189 (14.9)/12671.51 (1.14–1.99) 1.19 (0.90–1.58)
Injecting drug user104 (21.9)/4752.41 (1.87–3.10) 1.86 (1.20–2.89)
Other/unknown9 (7.8)/1150.73 (0.39–1.37) 0.91 (0.46–1.82)
CD4 count at enrolment  <0.001 < 0.001
≤ 50 cells/μL8 (1.6)/5041 1
51–200 cells/μL32 (3.7)/8622.39 (0.76–7.52) 2.24 (0.71–7.05)
201–350 cells/μL119 (13.5)/8839.66 (3.17–29.41) 8.49 (2.69–26.83)
351–500 cells/μL142 (31.8)/44628.96 (9.45–88.70) 27.58 (8.73–87.14)
>500 cells/μL125 (35.6)/35134.29 (12.03–97.73) 34.13 (11.59–100.48)
Viral load at enrolment  <0.001 < 0.001
< 4.0 log10 copies/mL102 (20.3)/5021 1
4.0–5.0 log10 copies/mL232 (18.8)/12310.91 (0.74–1.12) 1.11 (0.87–1.40)
> 5.0 log10 copies/mL92 (7)/13130.30 (0.20–0.43) 0.57 (0.40–0.80)
Hepatitis C  <0.001 0.0467
No285 (12.3)/23141 1
Yes129 (20.8)/6191.87 (1.56–2.26) 1.54 (1.09–2.18)
Unknown12 (10.6)/1130.85 (0.46–1.57) 0.85 (0.47–1.54)
Date when treatment criteria were fulfilled  0.014 < 0.001
2009–201075 (16.2)/4631 1
2007–2008138 (11.9)/11630.91 (0.68–1.23) 1.02 (0.72–1.44)
2004–2006213 (15)/14200.70 (0.52–0.93) 1.77 (1.31–2.41)

Outcome after treatment initiation

For the analysis of prognosis after treatment initiation, 2734 and 2532 patients had information on virological and immunological outcomes 1 year after treatment initiation, respectively, and 4301 patients had information on follow-up visits and vital status for the assessment of mortality after treatment initiation. The use of a nonrecommended regimen was associated with a lower probability of having a virological response, but it was not associated with immunological response. The association of the use of a nonrecommended regimen with virological and immunological response did not change after adjusting for all other factors (Table 5).

Table 5. Outcome for patients receiving nonrecommended treatments compared with patients receiving recommended/alternative treatments
OutcomePatients with response [n (%)/total]Univariable OR (95% CI)PMultivariable OR (95% CI)*P
  1. CI, confidence interval; HR, hazard ratio; OR, odds ratio.
  2. *Adjusted by age, gender, route of transmission, CD4 count and viral load at treatment initiation, hepatitis B or C virus infection, country of origin, education, number of hospital beds and date of treatment initiation.
Virological response (n = 2734)  0.019 0.058
Recommended2000 (80.1)/24711 1
Nonrecommended193 (73.4)/2630.65 (0.45-0.93) 0.68 (0.46-1.01)
Immunological response (n = 2532)  0.273 0.164
Recommended1705 (74.5)/22881 1
Nonrecommended177 (72.5)/2440.90 (0.75-1.08) 1.11 (0.95-1.29)
 Rate (deaths per 100 000 person-years) (95% CI)HRPHR*P
Mortality (n = 4301)  0.035 0.002
Recommended4.45 (3.81–5.21)1 1
Nonrecommended6.38 (4.43–9.18)1.61 (1.03–2.52) 1.65 (1.20–2.26)

After treatment initiation, 185 (4.1%) patients died: there were 29 deaths among patients receiving a nonrecommended treatment and 156 among those with recommended/alternative treatments. The use of a nonrecommended regimen was significantly associated with mortality [adjusted hazard ratio (HR) 1.64; 95% CI 1.20-2.26; P = 0.002] (Table 5).

The cause of death was known in 27 patients receiving nonrecommended treatments and 133 patients receiving recommended/alternative treatments. Among these patients, death was attributable to AIDS-defining illnesses in 12 patients (44.4%) and non-AIDS-related causes in 15 (55.6%) patients receiving nonrecommended regimens, and 78 (58.6%) and 55 (41.4%) patients receiving recommended/alternative regimens, respectively. There was no significant difference in the proportions of deaths attributable to AIDS-defining illnesses and non-AIDS-related causes between patients receiving recommended/alternative and nonrecommended treatments (P = 0.205).

When ‘later recommended’ regimens were considered as recommended/alternative, the association of nonrecommended treatments with mortality and virological and immunological responses did not change (data not shown).

Discussion

Our study found that compliance with the ‘what to start with’ and ‘when to start’ recommendations of Spanish national guidelines was high. The use of nonrecommended regimens was more likely in patients with CD4 counts > 500 cells/μl, HBV infection, and treatment initiation in the years 2004–2006 and in small hospitals. Delayed initiation of treatment was associated with higher CD4 counts at enrolment, HCV infection, use of injecting drugs and initiation during the years 2004–2006, and it was less likely in patients with higher viral loads. Nonrecommended regimens were associated with higher mortality and lack of virological response.

We found that the choice of initial antiretroviral regimen was appropriate in most cases: 91.5% of patients started treatment that was considered recommended or alternative by the guidelines. This is consistent with a study in the Spanish Asociación Médica Vach de Estudios Multicéntricos (VACH) cohort in the years 2004–2006, which found that initial treatment regimens were compliant with the guidelines in 95% of cases [8]. Similar proportions have been found in other European studies [6, 7, 12], but studies from the USA have found lower percentages of compliance [13-15]. These differences across studies can partly be attributed to different populations and health care systems, use of different guidelines and general differences in the experience of caring for HIV-infected patients.

Our cohort includes 28 centres from 13 of the 17 Spanish autonomous regions and our patients' characteristics are similar to those of patients included in the Spanish National HIV registry [20, 33], which supports the representativeness of our data for the Spanish territory. HIV treatment in Spain is provided only in specialized hospital-based reference units, which probably partly accounts for the high compliance with the national guidelines. However, there is a possibility that compliance might be slightly lower in other Spanish centres, as many of the hospitals included in our cohort are involved in the development of the guidelines.

The use of nonrecommended treatments and delayed start of treatment were more frequent during the years 2004–2006, which suggests increased acceptance of the guidelines and increased familiarity with treatment recommendations over time. The fact that smaller hospitals had a higher use of nonrecommended treatments also suggests that these centres might have less familiarity or knowledge of the guidelines. We found a higher risk of using nonrecommended treatments in patients with higher CD4 counts, which was also found in a recent Swiss study [7].

The higher risk of having a nonrecommended treatment in patients with hepatitis B might be a result of clinicians choosing nonrecommended regimens to avoid hepatotoxicity; however, we did not find similar results for patients with hepatitis C. Tenofovir, which is effective against HBV, was not considered a recommended treatment before November 2004, and it was prescribed in most of the patients with HBV infection during that time. To evaluate whether the use of tenofovir could have influenced our results for patients with hepatitis B, we performed a stratified analysis dividing the patients before and after November 2004: hepatitis B was still significantly associated with having a nonrecommended regimen during both periods (data not shown).

Almost one-third of patients starting a nonrecommended treatment did so with a treatment that was later recommended in further updates of the guidelines. This ‘anticipation’ shows that physicians' treatment decisions may be made based on the available evidence before new treatment advances can be incorporated in the guidelines. However, this did not seem to have an effect on our findings: we repeated all our analyses considering these treatments first as ‘nonrecommended’ and secondly as ‘recommended/alternative’, and none of our results were significantly changed.

The analysis of delayed start of treatment showed that higher CD4 counts at enrolment were very strongly associated with late treatment, suggesting that, although the guidelines are increasingly adding other risk factors as criteria for starting treatment (such as hepatitis C or B, or advanced age), clinicians may still base the decision to start treatment mainly on the CD4 count. Higher viral load, which was also added as a criterion for treatment initiation in later years, seemed to be associated with earlier start of treatment.

IDUs were more likely to have a delayed start of treatment. This finding has been reported previously [6, 34] and is probably attributable to concerns about low adherence in this group. As found in another study [6], hepatitis C was also associated with delayed treatment: this might seem surprising as antiretroviral treatment is recommended in patients with hepatitis C to avoid progression of liver damage. However, the recommendation for antiretroviral therapy in patients with hepatitis C was only introduced into the guidelines after February 2009 and therefore might not yet have had an effect on our patients. To account for the fact that hepatitis C and use of injecting drugs are usually correlated, we performed the same analysis stratified by transmission category: the effect of hepatitis C on treatment delay was no longer significant in IDUs, but it remained significant in all other categories (data not shown). This suggests that hepatitis C does not have an influence on the decision to start treatment among IDUs.

HIV nephropathy, cirrhosis and high cardiovascular risk were introduced as factors to be used as criteria for recommending treatment among patients with CD4 counts of 350−500 cells/μl in February 2009. Unfortunately we did not have information on these factors; however, we do not believe that such information would have greatly influenced our analysis of time to treatment initiation, as the number of patients with HIV nephropathy or cirrhosis caused by nonviral diseases would have been very small, and the number of patients fulfilling treatment criteria in that time period was low.

We defined ‘delayed treatment initiation’ as starting treatment more than 1 year after treatment was indicated (or not initiating it at all). This cut-off point was selected in order to ensure that all the patients analysed had a true delay of treatment. However, we also performed the analysis using a cut-off point of 6 months: our results were not changed and the risk factors for delayed treatment initiation were the same, except that the association with the date of treatment indication was no longer significant (data not shown).

The use of nonrecommended regimens was associated with a lack of virological response, but not with immunological response. Only two other studies have assessed the association of treatment guidelines with immunovirological response. One of them also found that treatments not in accordance with the guidelines were associated with poorer virological response, but there was no association with immunological response [7]. The second study found that patients receiving recommended treatments had a better virological and immunological response [13].

To our knowledge, this is the first study to find an association between compliance with clinical treatment guidelines and mortality. Patients receiving nonrecommended treatments had a significantly higher mortality rate compared with those receiving alternative or recommended treatments. This association must be interpreted cautiously, as it could be a result of residual confounding rather than causality. Unrecorded comorbid conditions, a higher proportion of primary resistance to antiretrovirals, or even lifestyle factors could have influenced both the choice of a nonrecommended treatment and the higher mortality risk. In fact, among those patients for whom information on resistance was available, there seemed to be some evidence of a higher proportion of primary resistance to antiretrovirals in patients receiving nonrecommended regimens, which did not reach statistical significance, probably because of the low number of cases. Also, the possibility of residual confounding as a result of unrecorded comorbid conditions may be suggested by the fact that patients receiving nonrecommended regimens seemed to have a lower proportion of AIDS-related deaths (which are more directly related to treatment efficacy) and a higher proportion of non-AIDS-related deaths (which are more related to comorbidity and lifestyle factors) compared with those receiving recommended/alternative regimens. However, because of the low number of cases, the differences between AIDS-related and non-AIDS-related deaths were not statistically significant.

Our study has several limitations. We had no information on which patients might have started treatment as part of a clinical trial: participation in clinical trials might in part explain the use of some nonrecommended treatments. We had no information on pregnancy status; however, the influence of pregnancy on our results was probably very small, because of the low proportion of women in our study, and also because zidovudine, which is prescribed as a first option in pregnancy, was a recommended or alternative treatment throughout the whole study period. Also, as our analysis was based on routine data, we had no information on the reasons why physicians chose to prescribe specific treatments or to deviate from the guidelines' recommendations. We also do not have information on whether the results of primary resistance testing were available to clinicians before starting treatment and thus we cannot determine whether these results guided treatment decisions.

In conclusion, in our multicentre cohort, compliance with national guidelines for antiretroviral treatment was high. It seems that the timing of treatment initiation was most strongly influenced by CD4 count and viral load, and that IDUs started treatment later than recommended, which was probably a consequence of concerns about lack of adherence. Despite the limitations inherent in the analysis of causality in a cohort study, the improved outcome of patients receiving recommended treatments suggests that it is worth encouraging compliance with clinical guidelines for antiretroviral treatment. Their acceptance increases with time as clinicians become more familiar with the guidelines, and more effort should probably be made to implement them in smaller centres that might have less experience in their use.

Acknowledgements

This study would not have been possible without the collaboration of all the patients, medical and nursery staff and data managers who have taken part in the project.

Conflicts of interest: None of the authors has any conflict of interest regarding this article.

Financial disclosure: CoRIS is funded by the Instituto de Salud Carlos III through the Red Temática de Investigación Cooperativa en Sida (RIS C03/173, RD 06/006). This study was funded by the Spanish Ministry of Health (Ayudas a la Investigación Independiente 2011, EC 11-443).

Contributions to authorship: All authors were involved in the setting up of the cohort and contributed to the design. JdA, PS-V, IS-G and SM asked the research question presented in this paper. All authors were involved in data collection. IS-G was responsible for statistical analyses with support from PS-V and JdA. IS-G wrote the first draft of the paper, which was supervised by PS-V, JdA and SM. All authors were involved in interpretation of the data and commented on interim drafts. All authors have read and approved the final draft.

Appendix: Appendix: CoRIS

Steering committee: Juan Berenguer, Julia del Amo, Federico García, Félix Gutiérrez, Pablo Labarga, Santiago Moreno and María Ángeles Muñoz.

Field work, analyses and writing committee: Ana María Caro-Murillo, Paz Sobrino Vegas, Santiago Pérez-Cachafeiro, Victoria Hernando Sebastián, Belén Alejos Ferreras, Débora Álvarez, Susana Monge, Inmaculada Jarrín and Mónica Trastoy.

BioBank: M. Ángeles Muñoz-Fernández, Isabel García-Merino, Coral Gómez Rico, Jorge Gallego de la Fuente and Almudena García Torre.

Participating centres

Hospital General Universitario de Alicante (Alicante): Joaquín Portilla Sogorb, Esperanza Merino de Lucas, Sergio Reus Bañuls, Vicente Boix Martínez, Livia Giner Oncina, Carmen Gadea Pastor, Irene Portilla Tamarit and Patricia Arcaina Toledo. Hospital Universitario de Canarias (Santa Cruz de Tenerife): Juan Luis Gómez Sirvent, Patricia Rodríguez Fortúnez, María Remedios Alemán Valls, María del Mar Alonso Socas, Ana María López Lirola, María Inmaculada Hernández Hernández and Felicitas Díaz-Flores. Hospital Carlos III (Madrid): Vicente Soriano, Pablo Labarga, Pablo Barreiro, Carol Castañares, Pablo Rivas, Andrés Ruiz, Francisco Blanco, Pilar García and Mercedes de Diego. Hospital Universitario Central de Asturias (Oviedo): Victor Asensi, Eulalia Valle and José Antonio Cartón. Hospital Clinic (Barcelona): José M. Miró, María López-Dieguez, Christian Manzardo, Laura Zamora, Iñaki Pérez, Mª Teresa García, Carmen Ligero, José Luis Blanco, Felipe García-Alcaide, Esteban Martínez, Josep Mallolas and José M. Gatell. Hospital Doce de Octubre (Madrid): Rafael Rubio, Federico Pulido, Silvana Fiorante, Jara Llenas, Violeta Rodríguez and Mariano Matarranz. Hospital Donostia (San Sebastián): José Antonio Iribarren, Julio Arrizabalaga, María José Aramburu, Xabier Camino, Francisco Rodríguez-Arrondo, Miguel Ángel von Wichmann, Lidia Pascual Tomé, Miguel Ángel Goenaga, Mª Jesús Bustinduy and Harkaitz Azkune Galparsoro. Hospital General Universitario de Elche (Elche): Félix Gutiérrez, Mar Masiá, José Manuel Ramos, Sergio Padilla, Andrés Navarro, Fernando Montolio, Yolanda Peral and Catalina Robledano García. Hospital Germans Trías i Pujol (Badalona): Bonaventura Clotet, Cristina Tural, Lidia Ruiz, Cristina Miranda, Roberto Muga, Jordi Tor and Arantza Sanvisens. Hospital Gregorio Marañón (Madrid): Juan Berenguer, Juan Carlos López Bernaldo de Quirós, Pilar Miralles, Jaime Cosín Ochaíta, Matilde Sánchez Conde, Isabel Gutiérrez Cuellar, Margarita Ramírez Schacke, Belén Padilla Ortega and Paloma Gijón Vidaurreta. Hospital Universitari de Tarragona Joan XXIII, IISPV, Universitat Rovira i Virgili (Tarragona): Francesc Vidal, Joaquín Peraire, Consuelo Viladés, Sergio Veloso, Montserrat Vargas, Miguel López-Dupla, Montserrat Olona, Alba Aguilar, Joan Joseph Sirvent, Antoni Soriano and Rami A. A. Qaneta. Hospital Universitario La Fe (Valencia): José López Aldeguer, Marino Blanes Juliá, José Lacruz Rodrigo, Miguel Salavert, Marta Montero, Eva Calabuig and Sandra Cuéllar. Hospital Universitário La Paz (Madrid): Juan González García, Ignacio Bernardino de la Serna, José María Peña Sánchez de Rivera, Marta Mora Rillo, José Ramón Arribas López, María Luisa Montes Ramírez, José Francisco Pascual Pareja, Blanca Arribas, Juan Miguel Castro, Fco Javier Zamora Vargas and Ignacio Pérez Valero. Hospital de la Princesa (Madrid): Ignacio de los Santos, Jesús Sanz Sanz, Johana Rodríguez, Ana Salas Aparicio and Cristina Sarriá Cepeda. Hospital San Pedro-CIBIR (Logroño): José Antonio Oteo, José Ramón Blanco, Valvanera Ibarra, Luis Metola, Mercedes Sanz and Laura Pérez-Martínez. Hospital San Pedro II (Logroño): Javier Pinilla Moraza. Hospital Universitario Mutua de Terrassa (Terrassa): David Dalmau, Angels Jaén Manzanera, Mireia Cairó Llobell, Daniel Irigoyen Puig, Laura Ibáñez, Queralt Jordano Montañez, Mariona Xercavins Valls, Javier Martinez-Lacasa, Pablo Velli and Roser Font. Hospital de Navarra (Pamplona): Julio Sola Boneta, Javier Uriz, Jesús Castiello, Jesús Reparaz, María Jesús Arraiza, Carmen Irigoyen and David Mozas. Hospital Parc Taulí (Sabadell): Ferrán Segura, María José Amengual, Eva Penelo, Gemma Navarro, Montserrat Sala, Manuel Cervantes and Valentín Pineda. Hospital Ramón y Cajal (Madrid): Santiago Moreno, José Luis Casado, Fernando Dronda, Ana Moreno, María Jesús Pérez Elías, Dolores López, Carolina Gutiérrez, Beatriz Hernández, María Pumares and Paloma Martí. Hospital Reina Sofía (Murcia): Alfredo Cano Sánchez, Enrique Bernal Morell and Ángeles Muñoz Pérez. Hospital San Cecilio (Granada): Federico García García, José Hernández Quero, Alejandro Peña Monje, Leopoldo Muñoz Medina and Jorge Parra Ruiz. Centro Sanitario Sandoval (Madrid): Jorge Del Romero Guerrero, Carmen Rodríguez Martín, Teresa Puerta López, Juan Carlos Carrió Montiel and Cristina González. Hospital Universitario Santiago de Compostela (Santiago de Compostela): Antonio Antela, Arturo Prieto and Elena Losada. Hospital Son Espases (Palma de Mallorca): Melchor Riera, Javier Murillas, Maria Peñaranda, Maria Leyes, Mª Angels Ribas, Antoni Campins, Concepcion Villalonga and Carmen Vidal. Hospital Universitario de Valme (Sevilla): Juan Antonio Pineda, Eva Recio Sánchez, Fernando Lozano de León, Juan Macías, José del Valle, Jesús Gómez-Mateos and Rosario Mata. Hospital Virgen de la Victoria (Málaga): Jesús Santos González, Manuel Márquez Solero, Isabel Viciana Ramos and Rosario Palacios Muñoz. Hospital Universitario Virgen del Rocío (Sevilla): Pompeyo Viciana, Manuel Leal, Luis Fernando López-Cortés and Mónica Trastoy.

Footnotes

  1. 3TC, lamivudine; ABC, abacavir; ZDV, zidovudine; APV/r, amprenavir/ritonavir; ATV, atazanavir; ATV/r, atazanavir/ritonavir; d4T, stavudine; ddC, zalcitabine; ddI, didanosine; DRV/r, darunavir; EFV, efavirenz; FOS, fosamprenavir; FOS/r, fosamprenavir/ritonavir; FTC, emtricitabine; IDV, indinavir; IDV/r, indinavir/ritonavir; LPV/r, lopinavir/ritonavir; NFV, nelfinavir; NRTI, nucleoside reverse transcriptase inhibitor; NVP, nevirapine; RTV, ritonavir; SQV/r, saquinavir/ritonavir; TDF, tenofovir.
  2. *No recommendations for alternative treatments were issued for the period January 2004 to October 2004.

Ancillary