Our findings reveal that 1-month-abstinent alcohol-dependents exhibit (1) an increase in the corticotropin amplitude and average baseline by SBPP but similar cortisol pulsatile characteristics and (2) a significantly blunted glucocorticoid response after oCRH stimulation in the presence of an corticotropin response not significantly different from that of the control group. These findings support a decrease in glucocorticoid responsiveness relative to corticotropin release in abstinent alcohol-dependent patients. In contrast, the corticotropin and cortisol responses to naloxone did not significantly differ between the two groups.
Alcohol-dependent patients revealed a relative increase in 0800 to 2000 hr corticotropin amplitude and baseline concentrations (by SPBB) and mean plasma concentrations in the presence of cortisol pulsatile characteristic and plasma concentrations within the control range. In part 1 of this study (Adinoff et al., 2005
), we reported that cortisol amplitude and mean concentrations were decreased in alcohol-dependent patients compared with controls from 2000 to 0800 hr, whereas corticotropin measures were similar. Both sets of findings suggest a decrease in cortisol responsiveness relative to corticotropin stimulation. Although two different alcohol-dependent patient samples were used for parts 1 and 2, these two groups were similar in age [t
(20) = 0.34, p
= 0.74], years of alcohol problems [t
(20) = 1.30, p
= 0.21], drinks over the previous 30 days [t
(19) = 1.03, p
= 0.31], and DrInC scores [t
(19) = 0.91, p
= 0.37]. Six controls were used for both studies (after at least a 6-week interval between parts 1 and 2). In addition, the leading and terminal points (0800 and 2000 hr) of the two 12-hr measures were in relatively close proximity.
Unlike the current study, Adinoff et al. (1991)
, Iranmanesh et al. (1989)
, and Loosen et al. (1991)
did not observe differences in the pulsatile characteristics of corticotropin or cortisol during abstinence in alcohol-dependent patients compared with controls. The SBPP modeling technique may have uncovered group differences in corticotropin amplitude not observed by others. The careful selection of our patient population, including the exclusion of patients with a lifetime history of other psychiatric disorders and/or a recent history of drug dependence and the specific 1-month time frame of abstinence, may also explain our findings. However, our patient data showed a fair amount of variability, and comparisons met only relatively low levels of statistical significance or trend levels.
Previous studies of the pituitary responsiveness in alcohol-dependent participants have consistently demonstrated a diminished response to various physical and psychological stressors. The specific time after abstinence, however, seems critical in understanding this literature. For example, Adinoff et al. (1990)
assessed corticotropin and glucocorticoid responsiveness to 1 μ
g/kg oCRH in alcohol-dependent men at 1 week, 3 weeks, 3 weeks to 6 months, and more than 6 months of abstinence. A marked suppression of the corticotropin response was observed at 1 week of abstinence compared with controls. Alcohol-dependent participants at 3 weeks of abstinence showed a significant increase in their corticotropin response compared with 1 week, nevertheless still significantly less than controls. The group that was abstinent for more than 3 weeks to 6 months revealed a corticotropin response similar to that of controls. This was similar to the work of Costa et al. (1996)
, who showed a blunted corticotropin and cortisol response to oCRH in 1-week to 2-weeks-abstinent subjects. Other investigators who showed a blunted corticotropin response to oCRH or human CRH studied subjects with varied periods of abstinence (Inder et al., 1995
: 10 days to 6 weeks; Loosen et al., 1993
: 15 to 39 days; Bardeleben et al., 1989
: 2 to 6 weeks), thus possibly mixing early abstinent (e.g., 3 weeks or fewer) corticotropin-suppressed subjects with later abstinent (e.g., more than 3 weeks) corticotropin-normalized subjects. In confirmation, Bailly et al. (1989)
did not describe a blunted corticotropin response in alcohol-dependent subjects assessed at 4 weeks of abstinence, although a blunted cortisol response was observed. However, the mean abstinent periods in the studies of Inder et al. (1995)
(25 ± 3 days) and Loosen et al. (1993)
(42.9 ± 25.7 days) were similar to ours (29.6 ± 1.0 days), and Ehrenreich et al. (1997)
showed a blunted corticotropin response to oCRH in 12-weeks-abstinent subjects. Other differences between our study and previous studies of the HPA axis in alcohol-dependent subjects include our rigorous exclusion criteria, including any other substance use disorders in the previous year (other than nicotine) and a lifetime history of axis I disorders. To our knowledge, the other studies referenced did not incorporate such stringent diagnostic criteria.
In our study, the absence of a decreased pituitary corticotropin response to oCRH in alcohol-dependent participants may also be related to the relatively low dose of CRH administered. The lower dose of CRH was used to avoid a ceiling glucocorticoid response to oCRH, thus potentially masking diminished adrenocortical reactivity in the alcohol-dependent group. Therefore, the current study administered an oCRH bolus of 0.4 μ
g/kg, compared with 1.0 μ
g/kg in Adinoff et al. (1990)
, Costa et al. (1996)
, Inder et al. (1995)
, and Loosen et al. (1993)
; 100 μ
g in Bardeleben et al. (1989)
, and 150 μ
g in Ehrenreich et al. (1997)
. The higher oCRH doses initially provoke a larger corticotropin and cortisol response, presumably inducing a more robust glucocorticoid inhibitory effect on the pituitary corticotrophs and suppressing further corticotropin release. Such a mechanism of diminished corticotropin responsivity has been posited in depression (Gold and Chrousos, 2002
), where a heightened adrenal response is thought to quickly suppress pituitary corticotropin production and release. We have recently observed that 1-month-abstinent alcohol-dependent patients exhibit increased corticotropin suppression after high-dose dexamethasone (Adinoff et al., 2005
), suggesting increased sensitivity of the pituitary glucocorticoid receptors. A less robust adrenocorticoid response to the lower dose of oCRH (0.4 μ
g/kg) in our patient population may therefore not have reached levels necessary to suppress pituitary secretion. In fact, as observed in , the corticotropin response in alcohol-dependent patients tended to be greater (not smaller) than in controls. However, this finding did not reach significance, perhaps because of the small sample size and substantial variance.
The administration of naloxone also did not induce an attenuated pituitary response in alcohol-dependent patients, as previously reported by Inder et al. (1995)
. Again, a more variable length of abstinence, a higher dose of naloxone (20 mg), and/or less restrictive selection criteria (including the ongoing use of disulfiram in the patient population) of Inder et al. (1995)
may account for these differences. Although the diminished plasma cortisol response to oCRH in alcohol-dependent patients would also be expected after the naloxone challenge 2 days later, we did not observe a blunted adrenocorticoid response to naloxone. It seems that the absence of this finding may have been due to the somewhat higher corticotropin response to naloxone in the alcohol-dependent patients (). However, differences in the corticotropin response did not reach significance between groups, possibly because of the small sample size and variance in the data. The heightened corticotropin response to naloxone may reflect premorbid differences, as Wand et al. (2001)
has reported that the corticotropin response to naloxone is increased in the children of alcohol-dependent individuals. Also unexpected was the absence of a significant correlation between the net integrated corticotropin response/net integrated cortisol response to oCRH relative to the net integrated corticotropin response/net integrated cortisol response to naloxone (r
= 0.289). Although the preceding oCRH stimulation test may have altered the subsequent naloxone stimulation test, the 48-hr intervening interval would be expected to minimize any carryover effect. Unfortunately, we did not counterbalance the oCRH and naloxone infusions.
The naloxone and oCRH challenge studies presented here were accompanied by a series of other basal and stimulated assessments of HPA axis reactivity in alcohol-dependent subjects (Adinoff et al., 2005
). These studies revealed significantly attenuated basal mean concentrations and pulsatile amplitude of plasma cortisol from 2000 to 0800 hr, a blunted glucocorticoid response to low-dose cosyntropin after dexamethasone pretreatment, and a heightened pituitary response to dexamethasone negative feedback. In toto and in context with other studies referenced above, these studies suggest the following scenario: Central activation of the HPA axis by chronic drinking and repeated withdrawal states induce a prolonged state of hypercortisolemia. The HPA axis adapts to the persistently increased plasma cortisol concentrations by down-regulating the hypothalamic, pituitary, and adrenal responses. At 4 weeks of abstinence, increased sensitivity of the pituitary corticotrophs to glucocorticoid negative feedback and diminished adrenocortical reactivity persists, accompanied by normalization of the hypothalamic and pituitary response to higher-level stimulation.
Our design had several notable strengths. Patients and controls were similar in age and race. Although ages were significantly different between the two groups, the mean ages were within 5 years of each other and were all within an age range during which HPA axis functioning remains stable. Neither group had significant medical disorders and were not taking any psychotropic or other medications that are known to alter HPA axis functioning. Alcohol-dependent patients were assessed within a relatively constricted period of abstinence, had no lifetime history of other non–substance-related psychiatric disorders, had no other substance use disorders in the previous year or frequent drug use not meeting DSM-IV
criteria (except nicotine dependence), and reported extensive drinking histories. Controls had no personal or family history of substance use disorders (except nicotine) and reported no lifetime history of other axis I disorders. These findings may not extend to alcohol-dependent females. In addition, alcohol-dependent patients were smokers, and controls were not. However, a previous study has reported that cigarette smoking in abstinent alcohol-dependent subjects did not alter the pituitary-adrenal response to a pharmacological challenge (Anthenelli et al., 2001
Raison and Miller (2003)
have posited that “hypoadrenal states” may be due to decreased glucocorticoid bioavailability, attenuated glucocorticoid responsiveness, and/or accentuated inhibitory feedback on the pituitary corticotrophs. Such hypoadrenal states have been observed in atypical depression, chronic fatigue, post–traumatic stress disorder, and fibromyalgia (see reviews in Heim et al., 2000a
; Raison and Miller, 2003
) Although the clinical relevance of the hypoadrenal states remains speculative, decreased concentrations of glucocorticoids are associated with early trauma, depressive mood states, dysphoric arousal, and behavioral dyscontrol and inhibition (Heim et al., 2000b
; King et al., 1990
; Tennes and Kreye, 1985
; Vanyukov et al., 1993
; Virkkunen, 1985
; Yehuda, 2001
). With respect to substance use disorder, glucocorticoid administration increases drug intake in animal models (Goeders, 2002
; Piazza et al., 1993
), possibly through activation of mesolimbic dopaminergic neurotransmission (Barrot et al., 2000
; Piazza et al., 1996
). Conversely, Nash and Maickel (1988)
have demonstrated that the unrestrained increase (e.g., absence of glucocorticoids) of hypothalamic CRH and pituitary corticotropin release after stress is causally related to increased drinking, and the increased drinking behavior was suppressed by the administration of exogenous glucocorticoids. Human studies demonstrate that inhibition of glucocorticoid synthesis in methadone-maintained patients is associated with the increased use of cocaine (Kosten et al., 2002
), and Kiefer et al. (2002)
have presented preliminary data suggesting a relation between low basal concentrations of glucocorticoids and early relapse. The positive relation between HPA axis disinhibition by naltrexone and craving in alcohol-dependent subjects (O’Malley et al., 2002
) suggests that glucocorticoid compensation in abstinent subjects may have therapeutic relevance.
Our findings offer further evidence for the persistent attenuation of adrenocortical reactivity in alcohol-dependent men. The diminished response is also sensitive to low levels of stimulation and is present for at least 1 month after cessation of drinking. Future studies should explore whether glucocorticoid hyposensitivity response persists with even longer periods of abstinence and confirm whether this disruption in neuroendocrine functioning is associated with a heightened relapse risk.