This study reported a significant increase in HRQoL during the first 48 weeks on HAART, with the bulk of this increase occurring during the first 16 weeks on treatment. Improvement in HRQoL occurred across all core domains assessed, as well as the physical health summary and mental health summary.
This study therefore supports the findings of Stangl et al [18
] and Jelsma et al [19
] who both reported an increase in HRQoL within the first three months of therapy in similar patient populations. The dramatic increase in HRQoL during the first few weeks of HAART occurred over the time period when patients usually experience the most significant gains in health. The greatest decrease in viral load happens within the first few weeks of treatment and mortality and morbidity rates begin to fall after just a month on HAART [29
There have been few analyses dealing with declines in PHS and MHS scores. In fact, negative HRQoL is often overshadowed by the overwhelming positive impact of HAART on HRQoL, and is therefore not reported. This research showed that although there was a general improvement in HRQoL on HAART, up to a third of participants experienced a decline in HRQoL during the first 48 weeks of HAART. Twenty-three percent of participants reported a drop in PHS score and 34% reported a drop in MHS score.
The most significant predictors of negative PHS and MHS were baseline HRQoL score, baseline log viral load and baseline CD4 count. The association between higher baseline HRQoL score and negative HRQoL could have been due to the fact that patients with higher baseline scores had less room for improvement and were therefore more likely to regress to the mean.
Baseline log viral load was strongly associated with negative PHS. Participants with higher baseline log viral loads were less likely to report negative PHS than participants with lower baseline log viral loads. Similarly, baseline log viral load and baseline CD4 count were associated with negative MHS. Participants with higher baseline log viral loads or lower baseline CD4 counts were less likely to report negative MHS than participants with lower baseline log viral loads or higher baseline CD4 counts.
Participants with more advanced disease, characterised by higher baseline viral loads and lower baseline CD4 counts, were less likely to report a decline in HRQoL than those with earlier disease. So it was the relatively well patients entering into the programme who were at greatest risk of experiencing negative HRQoL. These associations could be explained by the negative impact of symptoms on HRQoL [10
]. Patients with more advanced disease are more likely to have a greater number or intensity of symptoms than patients with less advanced disease pre-HAART [10
], but once on HAART, these symptoms improve [11
Although there is great concern about the impact of drug toxicities on HRQoL, few studies have directly assessed this association. In the developed world, symptoms have been shown to impact negatively on both physical and mental HRQoL [10
]. However, the nature of the symptoms and whether they were attributed to the disease process or to HAART was not clear.
In the developing world, Jelsma et al [19
] concluded that possible side effects of HAART had a negligible impact on HRQoL. This conclusion was based on the overall increase in HRQoL demonstrated for the cohort and did not specifically address those patients who experienced a decline in HRQoL.
Few drug toxicities were recorded during the first 48 weeks on HAART. While participants who experienced drug toxicity had lower PHS scores than participants without a drug toxicity at all time points (most notably during the 32 to 48 week treatment interval), only three (27%) participants with toxicity reported an actual decline in physical HRQoL between pre-HAART and week 48.
Drug toxicities, especially those related to stavudine use, may have a negative impact on physical HRQoL at the time of the toxicity. They do not, however, reduce overall gains in HRQoL. Drug toxicities had little impact on mental HRQoL.
The greatest strength of this study is its longitudinal design. This allowed the assessment of the associations between various socio-demographic and clinical factors and HRQoL, and allowed for inferences to be made about causal relationships.
Limitations to the study included possible selection bias and information bias related to the use of the MOS-SF36 instrument. The loss of up to 50% of patients, who had incomplete HRQoL data, from the final analysis may have resulted in a selection bias towards the healthier section of the cohort. However, as there were no significant differences in demographic, baseline and on-treatment characteristics between patients with incomplete and complete HRQoL data, it is unlikely that this was the case. This did impact on the number of patients available for analysis though, and may have limited the ability of the study to detect significant associations.
The MOS-SF36 instrument is a generic HRQoL measure and, like all generic instruments, may not be sensitive enough to measure the more specific aspects of HRQoL impacted on by the HIV disease process. This could lead to either an underestimation or overestimation of HRQoL scores and, more important, to a change in HRQoL scores, thereby under-reporting or over-reporting the actual impact of HAART on HRQoL.
The MOS-SF36 instrument is also prone to ceiling effects, where substantial numbers of patients get the highest possible score for a domain. This would have made it difficult for the instrument to pick up changes in HRQoL at the upper end of the scale, and may have been a problem with increasing time on treatment. Ceiling effects could have lead to the underestimation of HRQoL gains.
This study reported on HRQoL during the first 48 weeks of HAART only. As HAART-related drug toxicities, especially those related to stavudine, increase with length of time on HAART [21
], this study may have under-reported the impact of drug toxicities on gains in HRQoL.
Furthermore, this study only considered drug toxicities that were severe enough to prompt a change in antiretroviral therapy. Less severe toxicities that may also have impacted negatively on HRQoL were not reported on. This could have led to an underestimation of the true negative impact of drug toxicities on HRQoL.