Diphenhydramine Increases Ventilatory Drive during Alfentanil Infusion

Eight nonsmoking healthy volunteers (three women and five men), ranging in age from 23 to 38 yr and in weight from 55 to 91 kg, consented to participate in this institutional review board-approved study. The volunteers abstained from using alcohol and caffeine for 24 h and took nothing by mouth for at least 8 h before the study began. To minimize the effect of auditory stimulation during ventilatory testing, the supine volunteers listened to quiet classical music via headphones. We monitored blood pressure, electric activity of the heart, and arterial oxygen saturation (S^pO²) with an Ohmeda (Louisville, CO) model 3700 oximeter (ear probe, fast mode). A Datex (Helsinki, Finland) Capnomac II, calibrated with three reference mixtures of carbon dioxide in oxygen, continuously monitored the end-tidal pressure of carbon dioxide (PET^CO2) and fractional concentration of oxygen in inspired gas (FI^O2) from a sampling port built into the face mask in front of the volunteers' mouths. Subjects breathed via mask from a to-and-fro circuit with variable carbon dioxide absorption. [2]By continuously monitoring PET^CO2on a breath-by-breath basis and adjusting the flow of gas through the carbon dioxide absorber accordingly, we kept PET^CO2constant despite changes in the volunteers' ventilatory patterns. A Hans-Rudolph (Kansas City, MO) 3700 heated pneumotachograph with a Validyne (Northridge, CA) DP45 differential pressure transducer and electronic integrator determined ventilatory volumes at body temperature and ambient pressure. Before each set of measurements, we performed a three-point volume calibration and linearity check using a Collins (Braintree, MA) 3200 3-1 super syringe. An analog to digital converter and computer recorded breath-by-breath measurements of minute ventilation (V^E), tidal volume (V^T), S^pO², and PET^CO2.

To determine the steady state ventilatory response to carbon dioxide, the volunteers breathed hyperoxic mixtures (FI^O2> 0.6) of oxygen in nitrogen with PET^CO2held constant at approximately 46 or 54 mmHg for alternate volunteers. After a 6-min equilibration period, we recorded the steady state V^Eand PET^CO2during 30 breaths. Then we measured the ventilatory response to hypoxia at the same carbon dioxide tension (46 or 54 mmHg) using the isocapnic rebreathing method. [3]After rapidly (30 s) decreasing the FI^O2within the breathing circuit to 0.21 and allowing subjects to equilibrate for 2 additional min, we discontinued oxygen inflow to the circuit. As metabolic oxygen consumption progressively decreased FI^O2, we continuously recorded V^E, PET^CO2, and S^pO². We adjusted the nitrogen inflow to maintain a constant rebreathing bag volume, and we varied the carbon dioxide absorber flow to maintain a constant PET^CO2despite changes in ventilation. When S^pO²reached approximately 75% (usually 5 - 7 min after discontinuing oxygen inflow), we terminated data collection and administered 100% oxygen for approximately 30 s. After a 10-min recovery period, we repeated the steady state and hypoxic ventilatory measurements at the alternate carbon dioxide tension (54 or 46 mmHg), whichever was not used previously. Thus, in odd-numbered volunteers, ventilatory measurements were performed first at PET^CO2[almost equal to] 46 mmHg and then at 54 mmHg, whereas in even-numbered volunteers, the sequence was reversed (Figure 1).

After these baseline measurements, we administered a graded alfentanil infusion at 0.6 [micro sign]g [middle dot] kg^-1[middle dot] min^-1for 5 min, 0.18 [micro sign]g [middle dot] kg^-1[middle dot] min^-1for 10 min, and 0.05 [micro sign]g [middle dot] kg^-1[middle dot] min^-1thereafter. We previously showed that this infusion regimen provides constant plasma alfentanil levels of approximately 10 ng/ml and a modest, consistent depression of the ventilatory responses to hypercapnia and hypoxia. [2]Then we determined the ventilatory response to carbon dioxide and hypoxia using the same sequence of carbon dioxide tensions as before. As the alfentanil infusion continued, the volunteers received a single dose of diphenhydramine (0.7 mg/kg given intravenously) over 15 s. After allowing 10 min for effect-site equilibration, we determined the ventilatory response during combined alfentanil and diphenhydramine, again using the same sequence of carbon dioxide tensions.

Because of breath-to-breath variability in respiratory variables, five-breath average values of V^E, V^T, PET^CO2, and percent S^pO²were used throughout the analysis. To determine the steady state ventilatory response to carbon dioxide, we used 12 of these five-breath averages as data points for the linear regressions of V^Eand V^Tversus PET^CO2; six of these points have PET^CO2[almost equal to] 46 mmHg, whereas the other six points have PET^CO2[almost equal to] 54 mmHg. The ventilatory response to isocapnic hypoxemia was determined during mild (PET^CO2[almost equal to] 46 mmHg) and moderate (PET^CO2[almost equal to] 54 mmHg) hypercapnia by least-squares linear regression of V^Eversus S^pO². Values of V^Eat S^pO²= 90% indicated the position of the hypoxic ventilatory response curve.

We analyzed the data using two-way (volunteers x study conditions) analysis of variance with significance set at P <or= to 0.05. If analysis of variance revealed overall significant differences between study conditions (baseline, alfentanil, alfentanil plus diphenhydramine), we used Fisher's protected least-significant difference test to isolate the source of the differences. Data are shown as mean +/- SD.

After receiving alfentanil, all volunteers were mildly sedated, responsive to verbal commands in a normal tone, and had mild ptosis or a "glazed" appearance of the eyes. This corresponds to a change from 5 to 4 on the Observers Assessment of Alertness/Sedation Scale. [4]The addition of diphenhydramine produced additional subjective sedation (i.e., volunteers reported feeling more drowsy); however, none of the volunteers was sedated enough to meet the criteria for an Observers Assessment of Alertness/Sedation Scale score of 3 (marked ptosis or spontaneous eye closure with response to loud verbal commands) after receiving diphenhydramine.

The slope of the ventilatory response to hypercapnia decreased from 1.08 +/- 0.38 to 0.79 +/- 0.36 l [middle dot] min^-1[middle dot] mmHg^-1(P < 0.05, Table 1) during alfentanil infusion; with the addition of diphenhydramine, the slope increased to 1.17 +/- 0.28 l [middle dot] min^-1[middle dot] mmHg^-1(P < 0.05). The V^Eat PET^CO2[almost equal to] 46 mmHg (V^E46) decreased from 12.1 +/- 3.7 to 9.7 +/- 3.6 l/min (P < 0.05, Figure 2) during alfentanil infusion. This resulted from a decrease in the V^Tat PET^CO2= 46 mmHg (V^T46) from 882 +/- 347 to 699 +/- 319 ml (P < 0.05) because respiratory rate was unaffected by alfentanil. After diphenhydramine, V^E46 was 9.9 +/- 2.9 l/min and V^T46 was 608 +/- 134 ml; these values did not differ significantly from those observed during alfentanil, but they remained significantly lower than baseline (P < 0.05). During alfentanil infusion, V (^E) 54 decreased from 21.3 +/- 4.8 to 16.6 +/- 4.7 l/min (P < 0.05); with diphenhydramine, this variable increased to 20.5 +/- 3.0 l/min (P < 0.05). The decrease in V^E54 was associated with a borderline-significant decrease in V^Tfrom 1,377 +/- 529 to 1,075 +/- 340 l/min (P = 0.08) and no significant change in respiratory rate. In contrast, the increase in V^E54 with diphenhydramine was associated with an increase in respiratory rate to 18 +/- 3.8 breaths/min (P < 0.05), whereas the increase in V^Tto 1,164 +/- 211 was not significant.

During mild hypercapnia (PET^CO2[almost equal to] 46 mmHg), the slope of the hypoxic ventilatory response did not change significantly between baseline, alfentanil, and alfentanil plus diphenhydramine trials (0.56 +/- 0.38, 0.46 +/- 0.27, 0.44 +/- 0.21 l [middle dot] min^-1[middle dot] %S^pO^2-1, respectively). Similarly, the computed V^Eat a S^pO²= 90% (V^E90) was not significantly affected by alfentanil or diphenhydramine (11.8 +/- 2.6, 10.1 +/- 2.1, 10.1 +/- 1.4 l/min for baseline, alfentanil, and alfentanil plus diphenhydramine, respectively). The PET^CO2was 45.9 +/- 0.7, 46.1 +/- 0.7, and 45.6 +/- 0.5 mmHg during baseline, alfentanil, and alfentanil plus diphenhydramine trials, ensuring that our measurements of the hypoxic ventilatory response were not confounded by systematic changes in PET (^CO)². Figure 3shows an example of an individual hypoxic response determination.

During moderate hypercapnia (PET^CO2[almost equal to] 54 mmHg), neither the decrease in slope of the hypoxic ventilatory response from 1.20 +/- 0.80 to 0.96 +/- 0.52 1 [middle dot] min^-1[middle dot] %S^pO^2-1during alfentanil nor the subsequent increase in slope to 1.05 +/- 0.73 l [middle dot] min^-1[middle dot] %S^pO^2-1during alfentanil plus diphenhydramine was significant (Figure 4). However, V^E90 decreased significantly from 30.5 +/- 9.7 to 23.1 +/- 6.9 l/min during alfentanil (P < 0.05). After diphenhydramine, the increase in V^E90 to 27.2 +/- 9.2 l/min just failed to achieve significance (P = 0.057). The PET^CO2was 54.2 +/- 0.4, 54.4 +/- 0.9, and 54.5 +/- 0.9 mmHg during baseline, alfentanil, and alfentanil plus diphenhydramine, again implying that our measurements of hypoxic ventilatory response were not confounded by systematic changes in PET^CO2.

We designed this study using a relatively low-dose alfentanil infusion that produced a significant depression in ventilatory drive but still allowed us to observe further diphenhydramine-induced ventilatory depression if it occurred. The appropriateness of our alfentanil dose was confirmed by our observation that the decrease in ventilatory response to carbon dioxide was similar to that seen after epidural or intramuscular fentanyl, [5,6]alfentanil, [7]or morphine. [8-10] 

We began the diphenhydramine plus alfentanil ventilatory measurements 10 min after diphenhydramine was administered. Although the time necessary for equilibration of diphenhydramine at its effect sites in the brain has not been reported, clinical experience in the treatment of dystonic reactions after phenothiazine administration suggests a rapid (within minutes) onset of action. [11]In addition, our volunteers seemed to reach a stable level of subjective sedation within about 5 min. Despite the sedative effect of diphenhydramine, its concomitant administration did not further depress ventilatory drive. In fact, diphenhydramine increased the slope of the ventilatory response to hypercarbia, restoring it to baseline levels, and did not further depress hypoxic ventilatory response. Each volunteer was studied on a single day; although this precluded the use of a placebo control, it avoided the problems associated with day-to-day variations in ventilatory drive, which may be especially important in premenopausal women.

Can the present data elucidate the mechanism of the observed changes in ventilatory drive? Typically, opioid-induced ventilatory depression results in a downward, parallel shift in the carbon dioxide response curve, with a modest depression of the slope. Reversal with an opioid antagonist results in complete reversal of this effect throughout the measured range of carbon dioxide tensions. [12]In contrast, we found that although diphenhydramine increased the slope of the carbon dioxide response, ventilation remained significantly depressed at PET^CO2= 46 mmHg. In addition, the alfentanil-induced decrease in computed ventilation at PET (^CO)²= 54 mmHg primarily resulted from a decrease in V^T, whereas the increase after diphenhydramine resulted primarily from an increase in respiratory rate. These two observations suggest that diphenhydramine has a nonantagonistic, additive effect on ventilatory drive that is most pronounced at higher carbon dioxide tensions. This corresponds with our previous observations that the respiratory stimulant effect of diphenhydramine is most apparent at elevated carbon dioxide tensions. [1]the distinction between an additive and an antagonistic effect on opioid-induced ventilatory depression is of practical significance. If diphenhydramine's effect is mediated by an additive mechanism, as suggested by the current results, we would expect only partial reversal of the ventilatory depressant effects of larger opioid doses.

The mechanism underlying the ventilatory stimulant effect of diphenhydramine observed in the current study is an enigma, particularly because decreased levels of consciousness typically augment opioid-induced ventilatory depression. [13]Within the central nervous system, histamine activation of H1 and H2 receptors increases production of cyclic adenosine monophosphate, [14]which would be expected to stimulate ventilation. [15]On this basis, H1 blockade by diphenhydramine should depress rather than augment ventilatory drive. Diphenhydramine also has central anticholinergic activity. Because ventilatory drive is proportional to acetylcholine activity in medullary respiratory centers, [16]the anticholinergic activity of diphenhydramine would be expected to further depress ventilatory drive. However, in certain circumstances, increased acetylcholine activity after physostigmine may be associated with ventilatory depression: Gesell [17]observed that physostigmine depresses phrenic nerve activity in dogs, whereas Berkenbosch et al. [18]described a physostigmine-induced decrease in both the central and the peripheral components of the carbon dioxide response slope in cats. In addition, Spaulding et al. [19]reported impaired recovery of ventilatory drive when physostigmine was administered to diazepam-sedated volunteers. Therefore, in some circumstances, muscarinic blockade might stimulate rather than depress ventilation. In fact, Burton et al. [20]found that selective M2 antagonism in the isolated neonatal rat brain stem stimulates phrenic nerve output. Conceivably, selective cholinergic blockade by diphenhydramine could account for the ventilatory stimulation we observed. This mechanism is also consistent with our observation that the effect of diphenhydramine was most pronounced during hypercarpnic conditions, suggesting a central rather than peripheral mechanism of ventilatory stimulation.

The effect of diphenhydramine on the hypoxic ventilatory response was less pronounced than its effect on carbon dioxide response. Although alfentanil significantly decreased V^E90, the increase in V^E90 after diphenhydramine did not achieve significance at P = 0.05 using a two-tailed test. However, power analysis revealed that the probability of dipenhydramine causing a further decrease of 10% or more in V^E90 is remote (P = 0.016). Hypoxic ventillatory response consists of a rapid, peripherally mediated stimulation via the carotid chemoroceptors and a slower, centrally mediated ventillatory depression. Because of the shape of the oxyhemoglobin dissociation curve, it took only 2 r 3 min for the S^pO²to decrease from 90% to 75% (Figure 3). Therefore, it is likely that our results primarily reflect the peripheral component, although some centrally mediated decrease may have occurred. Nonetheless, because hypoxic challenges were administered in the baseline, alfentanil, and alfentanil plus diphenydramine states, it is unlikely that differences in the contribution of central versus peripheral control mechanisms during the three study conditions significantly affected our results.

Diphenhydramine is prescribed widely as an antipruritic, an antiemetic, and a sedative in patients receiving systemic or neuraxial opioids. The current findings help to explain the long-standing record of the safety of diphenhydramine when administered to various patients for these indications.